Related
How do I create an empty DataFrame, then add rows, one by one?
I created an empty DataFrame:
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
Then I can add a new row at the end and fill a single field with:
df = df._set_value(index=len(df), col='qty1', value=10.0)
It works for only one field at a time. What is a better way to add new row to df?
You can use df.loc[i], where the row with index i will be what you specify it to be in the dataframe.
>>> import pandas as pd
>>> from numpy.random import randint
>>> df = pd.DataFrame(columns=['lib', 'qty1', 'qty2'])
>>> for i in range(5):
>>> df.loc[i] = ['name' + str(i)] + list(randint(10, size=2))
>>> df
lib qty1 qty2
0 name0 3 3
1 name1 2 4
2 name2 2 8
3 name3 2 1
4 name4 9 6
In case you can get all data for the data frame upfront, there is a much faster approach than appending to a data frame:
Create a list of dictionaries in which each dictionary corresponds to an input data row.
Create a data frame from this list.
I had a similar task for which appending to a data frame row by row took 30 min, and creating a data frame from a list of dictionaries completed within seconds.
rows_list = []
for row in input_rows:
dict1 = {}
# get input row in dictionary format
# key = col_name
dict1.update(blah..)
rows_list.append(dict1)
df = pd.DataFrame(rows_list)
In the case of adding a lot of rows to dataframe, I am interested in performance. So I tried the four most popular methods and checked their speed.
Performance
Using .append (NPE's answer)
Using .loc (fred's answer)
Using .loc with preallocating (FooBar's answer)
Using dict and create DataFrame in the end (ShikharDua's answer)
Runtime results (in seconds):
Approach
1000 rows
5000 rows
10 000 rows
.append
0.69
3.39
6.78
.loc without prealloc
0.74
3.90
8.35
.loc with prealloc
0.24
2.58
8.70
dict
0.012
0.046
0.084
So I use addition through the dictionary for myself.
Code:
import pandas as pd
import numpy as np
import time
del df1, df2, df3, df4
numOfRows = 1000
# append
startTime = time.perf_counter()
df1 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows-4):
df1 = df1.append( dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']), ignore_index=True)
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df1.shape)
# .loc w/o prealloc
startTime = time.perf_counter()
df2 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows):
df2.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df2.shape)
# .loc with prealloc
df3 = pd.DataFrame(index=np.arange(0, numOfRows), columns=['A', 'B', 'C', 'D', 'E'] )
startTime = time.perf_counter()
for i in range( 1,numOfRows):
df3.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df3.shape)
# dict
startTime = time.perf_counter()
row_list = []
for i in range (0,5):
row_list.append(dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']))
for i in range( 1,numOfRows-4):
dict1 = dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E'])
row_list.append(dict1)
df4 = pd.DataFrame(row_list, columns=['A','B','C','D','E'])
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df4.shape)
P.S.: I believe my realization isn't perfect, and maybe there is some optimization that could be done.
You could use pandas.concat(). For details and examples, see Merge, join, and concatenate.
For example:
def append_row(df, row):
return pd.concat([
df,
pd.DataFrame([row], columns=row.index)]
).reset_index(drop=True)
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
new_row = pd.Series({'lib':'A', 'qty1':1, 'qty2': 2})
df = append_row(df, new_row)
NEVER grow a DataFrame!
Yes, people have already explained that you should NEVER grow a DataFrame, and that you should append your data to a list and convert it to a DataFrame once at the end. But do you understand why?
Here are the most important reasons, taken from my post here.
It is always cheaper/faster to append to a list and create a DataFrame in one go.
Lists take up less memory and are a much lighter data structure to work with, append, and remove.
dtypes are automatically inferred for your data. On the flip side, creating an empty frame of NaNs will automatically make them object, which is bad.
An index is automatically created for you, instead of you having to take care to assign the correct index to the row you are appending.
This is The Right Way™ to accumulate your data
data = []
for a, b, c in some_function_that_yields_data():
data.append([a, b, c])
df = pd.DataFrame(data, columns=['A', 'B', 'C'])
These options are horrible
append or concat inside a loop
append and concat aren't inherently bad in isolation. The
problem starts when you iteratively call them inside a loop - this
results in quadratic memory usage.
# Creates empty DataFrame and appends
df = pd.DataFrame(columns=['A', 'B', 'C'])
for a, b, c in some_function_that_yields_data():
df = df.append({'A': i, 'B': b, 'C': c}, ignore_index=True)
# This is equally bad:
# df = pd.concat(
# [df, pd.Series({'A': i, 'B': b, 'C': c})],
# ignore_index=True)
Empty DataFrame of NaNs
Never create a DataFrame of NaNs as the columns are initialized with
object (slow, un-vectorizable dtype).
# Creates DataFrame of NaNs and overwrites values.
df = pd.DataFrame(columns=['A', 'B', 'C'], index=range(5))
for a, b, c in some_function_that_yields_data():
df.loc[len(df)] = [a, b, c]
The Proof is in the Pudding
Timing these methods is the fastest way to see just how much they differ in terms of their memory and utility.
Benchmarking code for reference.
It's posts like this that remind me why I'm a part of this community. People understand the importance of teaching folks getting the right answer with the right code, not the right answer with wrong code. Now you might argue that it is not an issue to use loc or append if you're only adding a single row to your DataFrame. However, people often look to this question to add more than just one row - often the requirement is to iteratively add a row inside a loop using data that comes from a function (see related question). In that case it is important to understand that iteratively growing a DataFrame is not a good idea.
If you know the number of entries ex ante, you should preallocate the space by also providing the index (taking the data example from a different answer):
import pandas as pd
import numpy as np
# we know we're gonna have 5 rows of data
numberOfRows = 5
# create dataframe
df = pd.DataFrame(index=np.arange(0, numberOfRows), columns=('lib', 'qty1', 'qty2') )
# now fill it up row by row
for x in np.arange(0, numberOfRows):
#loc or iloc both work here since the index is natural numbers
df.loc[x] = [np.random.randint(-1,1) for n in range(3)]
In[23]: df
Out[23]:
lib qty1 qty2
0 -1 -1 -1
1 0 0 0
2 -1 0 -1
3 0 -1 0
4 -1 0 0
Speed comparison
In[30]: %timeit tryThis() # function wrapper for this answer
In[31]: %timeit tryOther() # function wrapper without index (see, for example, #fred)
1000 loops, best of 3: 1.23 ms per loop
100 loops, best of 3: 2.31 ms per loop
And - as from the comments - with a size of 6000, the speed difference becomes even larger:
Increasing the size of the array (12) and the number of rows (500) makes
the speed difference more striking: 313ms vs 2.29s
mycolumns = ['A', 'B']
df = pd.DataFrame(columns=mycolumns)
rows = [[1,2],[3,4],[5,6]]
for row in rows:
df.loc[len(df)] = row
You can append a single row as a dictionary using the ignore_index option.
>>> f = pandas.DataFrame(data = {'Animal':['cow','horse'], 'Color':['blue', 'red']})
>>> f
Animal Color
0 cow blue
1 horse red
>>> f.append({'Animal':'mouse', 'Color':'black'}, ignore_index=True)
Animal Color
0 cow blue
1 horse red
2 mouse black
For efficient appending, see How to add an extra row to a pandas dataframe and Setting With Enlargement.
Add rows through loc/ix on non existing key index data. For example:
In [1]: se = pd.Series([1,2,3])
In [2]: se
Out[2]:
0 1
1 2
2 3
dtype: int64
In [3]: se[5] = 5.
In [4]: se
Out[4]:
0 1.0
1 2.0
2 3.0
5 5.0
dtype: float64
Or:
In [1]: dfi = pd.DataFrame(np.arange(6).reshape(3,2),
.....: columns=['A','B'])
.....:
In [2]: dfi
Out[2]:
A B
0 0 1
1 2 3
2 4 5
In [3]: dfi.loc[:,'C'] = dfi.loc[:,'A']
In [4]: dfi
Out[4]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
In [5]: dfi.loc[3] = 5
In [6]: dfi
Out[6]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
3 5 5 5
For the sake of a Pythonic way:
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
res = res.append([{'qty1':10.0}], ignore_index=True)
print(res.head())
lib qty1 qty2
0 NaN 10.0 NaN
You can also build up a list of lists and convert it to a dataframe -
import pandas as pd
columns = ['i','double','square']
rows = []
for i in range(6):
row = [i, i*2, i*i]
rows.append(row)
df = pd.DataFrame(rows, columns=columns)
giving
i double square
0 0 0 0
1 1 2 1
2 2 4 4
3 3 6 9
4 4 8 16
5 5 10 25
If you always want to add a new row at the end, use this:
df.loc[len(df)] = ['name5', 9, 0]
I figured out a simple and nice way:
>>> df
A B C
one 1 2 3
>>> df.loc["two"] = [4,5,6]
>>> df
A B C
one 1 2 3
two 4 5 6
Note the caveat with performance as noted in the comments.
This is not an answer to the OP question, but a toy example to illustrate ShikharDua's answer which I found very useful.
While this fragment is trivial, in the actual data I had 1,000s of rows, and many columns, and I wished to be able to group by different columns and then perform the statistics below for more than one target column. So having a reliable method for building the data frame one row at a time was a great convenience. Thank you ShikharDua!
import pandas as pd
BaseData = pd.DataFrame({ 'Customer' : ['Acme','Mega','Acme','Acme','Mega','Acme'],
'Territory' : ['West','East','South','West','East','South'],
'Product' : ['Econ','Luxe','Econ','Std','Std','Econ']})
BaseData
columns = ['Customer','Num Unique Products', 'List Unique Products']
rows_list=[]
for name, group in BaseData.groupby('Customer'):
RecordtoAdd={} #initialise an empty dict
RecordtoAdd.update({'Customer' : name}) #
RecordtoAdd.update({'Num Unique Products' : len(pd.unique(group['Product']))})
RecordtoAdd.update({'List Unique Products' : pd.unique(group['Product'])})
rows_list.append(RecordtoAdd)
AnalysedData = pd.DataFrame(rows_list)
print('Base Data : \n',BaseData,'\n\n Analysed Data : \n',AnalysedData)
You can use a generator object to create a Dataframe, which will be more memory efficient over the list.
num = 10
# Generator function to generate generator object
def numgen_func(num):
for i in range(num):
yield ('name_{}'.format(i), (i*i), (i*i*i))
# Generator expression to generate generator object (Only once data get populated, can not be re used)
numgen_expression = (('name_{}'.format(i), (i*i), (i*i*i)) for i in range(num) )
df = pd.DataFrame(data=numgen_func(num), columns=('lib', 'qty1', 'qty2'))
To add raw to existing DataFrame you can use append method.
df = df.append([{ 'lib': "name_20", 'qty1': 20, 'qty2': 400 }])
Instead of a list of dictionaries as in ShikharDua's answer (row-based), we can also represent our table as a dictionary of lists (column-based), where each list stores one column in row-order, given we know our columns beforehand. At the end we construct our DataFrame once.
In both cases, the dictionary keys are always the column names. Row order is stored implicitly as order in a list. For c columns and n rows, this uses one dictionary of c lists, versus one list of n dictionaries. The list-of-dictionaries method has each dictionary storing all keys redundantly and requires creating a new dictionary for every row. Here we only append to lists, which overall is the same time complexity (adding entries to list and dictionary are both amortized constant time) but may have less overhead due to being a simple operation.
# Current data
data = {"Animal":["cow", "horse"], "Color":["blue", "red"]}
# Adding a new row (be careful to ensure every column gets another value)
data["Animal"].append("mouse")
data["Color"].append("black")
# At the end, construct our DataFrame
df = pd.DataFrame(data)
# Animal Color
# 0 cow blue
# 1 horse red
# 2 mouse black
Create a new record (data frame) and add to old_data_frame.
Pass a list of values and the corresponding column names to create a new_record (data_frame):
new_record = pd.DataFrame([[0, 'abcd', 0, 1, 123]], columns=['a', 'b', 'c', 'd', 'e'])
old_data_frame = pd.concat([old_data_frame, new_record])
Here is the way to add/append a row in a Pandas DataFrame:
def add_row(df, row):
df.loc[-1] = row
df.index = df.index + 1
return df.sort_index()
add_row(df, [1,2,3])
It can be used to insert/append a row in an empty or populated Pandas DataFrame.
If you want to add a row at the end, append it as a list:
valuestoappend = [va1, val2, val3]
res = res.append(pd.Series(valuestoappend, index = ['lib', 'qty1', 'qty2']), ignore_index = True)
Another way to do it (probably not very performant):
# add a row
def add_row(df, row):
colnames = list(df.columns)
ncol = len(colnames)
assert ncol == len(row), "Length of row must be the same as width of DataFrame: %s" % row
return df.append(pd.DataFrame([row], columns=colnames))
You can also enhance the DataFrame class like this:
import pandas as pd
def add_row(self, row):
self.loc[len(self.index)] = row
pd.DataFrame.add_row = add_row
All you need is loc[df.shape[0]] or loc[len(df)]
# Assuming your df has 4 columns (str, int, str, bool)
df.loc[df.shape[0]] = ['col1Value', 100, 'col3Value', False]
or
df.loc[len(df)] = ['col1Value', 100, 'col3Value', False]
You can concatenate two DataFrames for this. I basically came across this problem to add a new row to an existing DataFrame with a character index (not numeric).
So, I input the data for a new row in a duct() and index in a list.
new_dict = {put input for new row here}
new_list = [put your index here]
new_df = pd.DataFrame(data=new_dict, index=new_list)
df = pd.concat([existing_df, new_df])
initial_data = {'lib': np.array([1,2,3,4]), 'qty1': [1,2,3,4], 'qty2': [1,2,3,4]}
df = pd.DataFrame(initial_data)
df
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
val_1 = [10]
val_2 = [14]
val_3 = [20]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
You can use a for loop to iterate through values or can add arrays of values.
val_1 = [10, 11, 12, 13]
val_2 = [14, 15, 16, 17]
val_3 = [20, 21, 22, 43]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
1 11 15 21
2 12 16 22
3 13 17 43
Make it simple. By taking a list as input which will be appended as a row in the data-frame:
import pandas as pd
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
for i in range(5):
res_list = list(map(int, input().split()))
res = res.append(pd.Series(res_list, index=['lib', 'qty1', 'qty2']), ignore_index=True)
pandas.DataFrame.append
DataFrame.append(self, other, ignore_index=False, verify_integrity=False, sort=False) → 'DataFrame'
Code
df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
df.append(df2)
With ignore_index set to True:
df.append(df2, ignore_index=True)
If you have a data frame df and want to add a list new_list as a new row to df, you can simply do:
df.loc[len(df)] = new_list
If you want to add a new data frame new_df under data frame df, then you can use:
df.append(new_df)
We often see the construct df.loc[subscript] = … to assign to one DataFrame row. Mikhail_Sam posted benchmarks containing, among others, this construct as well as the method using dict and create DataFrame in the end. He found the latter to be the fastest by far.
But if we replace the df3.loc[i] = … (with preallocated DataFrame) in his code with df3.values[i] = …, the outcome changes significantly, in that that method performs similar to the one using dict. So we should more often take the use of df.values[subscript] = … into consideration. However note that .values takes a zero-based subscript, which may be different from the DataFrame.index.
Before going to add a row, we have to convert the dataframe to a dictionary. There you can see the keys as columns in the dataframe and the values of the columns are again stored in the dictionary, but there the key for every column is the index number in the dataframe.
That idea makes me to write the below code.
df2 = df.to_dict()
values = ["s_101", "hyderabad", 10, 20, 16, 13, 15, 12, 12, 13, 25, 26, 25, 27, "good", "bad"] # This is the total row that we are going to add
i = 0
for x in df.columns: # Here df.columns gives us the main dictionary key
df2[x][101] = values[i] # Here the 101 is our index number. It is also the key of the sub dictionary
i += 1
If all data in your Dataframe has the same dtype you might use a NumPy array. You can write rows directly into the predefined array and convert it to a dataframe at the end.
It seems to be even faster than converting a list of dicts.
import pandas as pd
import numpy as np
from string import ascii_uppercase
startTime = time.perf_counter()
numcols, numrows = 5, 10000
npdf = np.ones((numrows, numcols))
for row in range(numrows):
npdf[row, 0:] = np.random.randint(0, 100, (1, numcols))
df5 = pd.DataFrame(npdf, columns=list(ascii_uppercase[:numcols]))
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df5.shape)
This code snippet uses a list of dictionaries to update the data frame. It adds on to ShikharDua's and Mikhail_Sam's answers.
import pandas as pd
colour = ["red", "big", "tasty"]
fruits = ["apple", "banana", "cherry"]
dict1={}
feat_list=[]
for x in colour:
for y in fruits:
# print(x, y)
dict1 = dict([('x',x),('y',y)])
# print(f'dict 1 {dict1}')
feat_list.append(dict1)
# print(f'feat_list {feat_list}')
feat_df=pd.DataFrame(feat_list)
feat_df.to_csv('feat1.csv')
I want to change the column labels of a Pandas DataFrame from
['$a', '$b', '$c', '$d', '$e']
to
['a', 'b', 'c', 'd', 'e']
Rename Specific Columns
Use the df.rename() function and refer the columns to be renamed. Not all the columns have to be renamed:
df = df.rename(columns={'oldName1': 'newName1', 'oldName2': 'newName2'})
# Or rename the existing DataFrame (rather than creating a copy)
df.rename(columns={'oldName1': 'newName1', 'oldName2': 'newName2'}, inplace=True)
Minimal Code Example
df = pd.DataFrame('x', index=range(3), columns=list('abcde'))
df
a b c d e
0 x x x x x
1 x x x x x
2 x x x x x
The following methods all work and produce the same output:
df2 = df.rename({'a': 'X', 'b': 'Y'}, axis=1) # new method
df2 = df.rename({'a': 'X', 'b': 'Y'}, axis='columns')
df2 = df.rename(columns={'a': 'X', 'b': 'Y'}) # old method
df2
X Y c d e
0 x x x x x
1 x x x x x
2 x x x x x
Remember to assign the result back, as the modification is not-inplace. Alternatively, specify inplace=True:
df.rename({'a': 'X', 'b': 'Y'}, axis=1, inplace=True)
df
X Y c d e
0 x x x x x
1 x x x x x
2 x x x x x
From v0.25, you can also specify errors='raise' to raise errors if an invalid column-to-rename is specified. See v0.25 rename() docs.
Reassign Column Headers
Use df.set_axis() with axis=1 and inplace=False (to return a copy).
df2 = df.set_axis(['V', 'W', 'X', 'Y', 'Z'], axis=1, inplace=False)
df2
V W X Y Z
0 x x x x x
1 x x x x x
2 x x x x x
This returns a copy, but you can modify the DataFrame in-place by setting inplace=True (this is the default behaviour for versions <=0.24 but is likely to change in the future).
You can also assign headers directly:
df.columns = ['V', 'W', 'X', 'Y', 'Z']
df
V W X Y Z
0 x x x x x
1 x x x x x
2 x x x x x
Just assign it to the .columns attribute:
>>> df = pd.DataFrame({'$a':[1,2], '$b': [10,20]})
>>> df
$a $b
0 1 10
1 2 20
>>> df.columns = ['a', 'b']
>>> df
a b
0 1 10
1 2 20
The rename method can take a function, for example:
In [11]: df.columns
Out[11]: Index([u'$a', u'$b', u'$c', u'$d', u'$e'], dtype=object)
In [12]: df.rename(columns=lambda x: x[1:], inplace=True)
In [13]: df.columns
Out[13]: Index([u'a', u'b', u'c', u'd', u'e'], dtype=object)
As documented in Working with text data:
df.columns = df.columns.str.replace('$', '')
Pandas 0.21+ Answer
There have been some significant updates to column renaming in version 0.21.
The rename method has added the axis parameter which may be set to columns or 1. This update makes this method match the rest of the pandas API. It still has the index and columns parameters but you are no longer forced to use them.
The set_axis method with the inplace set to False enables you to rename all the index or column labels with a list.
Examples for Pandas 0.21+
Construct sample DataFrame:
df = pd.DataFrame({'$a':[1,2], '$b': [3,4],
'$c':[5,6], '$d':[7,8],
'$e':[9,10]})
$a $b $c $d $e
0 1 3 5 7 9
1 2 4 6 8 10
Using rename with axis='columns' or axis=1
df.rename({'$a':'a', '$b':'b', '$c':'c', '$d':'d', '$e':'e'}, axis='columns')
or
df.rename({'$a':'a', '$b':'b', '$c':'c', '$d':'d', '$e':'e'}, axis=1)
Both result in the following:
a b c d e
0 1 3 5 7 9
1 2 4 6 8 10
It is still possible to use the old method signature:
df.rename(columns={'$a':'a', '$b':'b', '$c':'c', '$d':'d', '$e':'e'})
The rename function also accepts functions that will be applied to each column name.
df.rename(lambda x: x[1:], axis='columns')
or
df.rename(lambda x: x[1:], axis=1)
Using set_axis with a list and inplace=False
You can supply a list to the set_axis method that is equal in length to the number of columns (or index). Currently, inplace defaults to True, but inplace will be defaulted to False in future releases.
df.set_axis(['a', 'b', 'c', 'd', 'e'], axis='columns', inplace=False)
or
df.set_axis(['a', 'b', 'c', 'd', 'e'], axis=1, inplace=False)
Why not use df.columns = ['a', 'b', 'c', 'd', 'e']?
There is nothing wrong with assigning columns directly like this. It is a perfectly good solution.
The advantage of using set_axis is that it can be used as part of a method chain and that it returns a new copy of the DataFrame. Without it, you would have to store your intermediate steps of the chain to another variable before reassigning the columns.
# new for pandas 0.21+
df.some_method1()
.some_method2()
.set_axis()
.some_method3()
# old way
df1 = df.some_method1()
.some_method2()
df1.columns = columns
df1.some_method3()
Since you only want to remove the $ sign in all column names, you could just do:
df = df.rename(columns=lambda x: x.replace('$', ''))
OR
df.rename(columns=lambda x: x.replace('$', ''), inplace=True)
Renaming columns in Pandas is an easy task.
df.rename(columns={'$a': 'a', '$b': 'b', '$c': 'c', '$d': 'd', '$e': 'e'}, inplace=True)
df.columns = ['a', 'b', 'c', 'd', 'e']
It will replace the existing names with the names you provide, in the order you provide.
Use:
old_names = ['$a', '$b', '$c', '$d', '$e']
new_names = ['a', 'b', 'c', 'd', 'e']
df.rename(columns=dict(zip(old_names, new_names)), inplace=True)
This way you can manually edit the new_names as you wish. It works great when you need to rename only a few columns to correct misspellings, accents, remove special characters, etc.
One line or Pipeline solutions
I'll focus on two things:
OP clearly states
I have the edited column names stored it in a list, but I don't know how to replace the column names.
I do not want to solve the problem of how to replace '$' or strip the first character off of each column header. OP has already done this step. Instead I want to focus on replacing the existing columns object with a new one given a list of replacement column names.
df.columns = new where new is the list of new columns names is as simple as it gets. The drawback of this approach is that it requires editing the existing dataframe's columns attribute and it isn't done inline. I'll show a few ways to perform this via pipelining without editing the existing dataframe.
Setup 1
To focus on the need to rename of replace column names with a pre-existing list, I'll create a new sample dataframe df with initial column names and unrelated new column names.
df = pd.DataFrame({'Jack': [1, 2], 'Mahesh': [3, 4], 'Xin': [5, 6]})
new = ['x098', 'y765', 'z432']
df
Jack Mahesh Xin
0 1 3 5
1 2 4 6
Solution 1
pd.DataFrame.rename
It has been said already that if you had a dictionary mapping the old column names to new column names, you could use pd.DataFrame.rename.
d = {'Jack': 'x098', 'Mahesh': 'y765', 'Xin': 'z432'}
df.rename(columns=d)
x098 y765 z432
0 1 3 5
1 2 4 6
However, you can easily create that dictionary and include it in the call to rename. The following takes advantage of the fact that when iterating over df, we iterate over each column name.
# Given just a list of new column names
df.rename(columns=dict(zip(df, new)))
x098 y765 z432
0 1 3 5
1 2 4 6
This works great if your original column names are unique. But if they are not, then this breaks down.
Setup 2
Non-unique columns
df = pd.DataFrame(
[[1, 3, 5], [2, 4, 6]],
columns=['Mahesh', 'Mahesh', 'Xin']
)
new = ['x098', 'y765', 'z432']
df
Mahesh Mahesh Xin
0 1 3 5
1 2 4 6
Solution 2
pd.concat using the keys argument
First, notice what happens when we attempt to use solution 1:
df.rename(columns=dict(zip(df, new)))
y765 y765 z432
0 1 3 5
1 2 4 6
We didn't map the new list as the column names. We ended up repeating y765. Instead, we can use the keys argument of the pd.concat function while iterating through the columns of df.
pd.concat([c for _, c in df.items()], axis=1, keys=new)
x098 y765 z432
0 1 3 5
1 2 4 6
Solution 3
Reconstruct. This should only be used if you have a single dtype for all columns. Otherwise, you'll end up with dtype object for all columns and converting them back requires more dictionary work.
Single dtype
pd.DataFrame(df.values, df.index, new)
x098 y765 z432
0 1 3 5
1 2 4 6
Mixed dtype
pd.DataFrame(df.values, df.index, new).astype(dict(zip(new, df.dtypes)))
x098 y765 z432
0 1 3 5
1 2 4 6
Solution 4
This is a gimmicky trick with transpose and set_index. pd.DataFrame.set_index allows us to set an index inline, but there is no corresponding set_columns. So we can transpose, then set_index, and transpose back. However, the same single dtype versus mixed dtype caveat from solution 3 applies here.
Single dtype
df.T.set_index(np.asarray(new)).T
x098 y765 z432
0 1 3 5
1 2 4 6
Mixed dtype
df.T.set_index(np.asarray(new)).T.astype(dict(zip(new, df.dtypes)))
x098 y765 z432
0 1 3 5
1 2 4 6
Solution 5
Use a lambda in pd.DataFrame.rename that cycles through each element of new.
In this solution, we pass a lambda that takes x but then ignores it. It also takes a y but doesn't expect it. Instead, an iterator is given as a default value and I can then use that to cycle through one at a time without regard to what the value of x is.
df.rename(columns=lambda x, y=iter(new): next(y))
x098 y765 z432
0 1 3 5
1 2 4 6
And as pointed out to me by the folks in sopython chat, if I add a * in between x and y, I can protect my y variable. Though, in this context I don't believe it needs protecting. It is still worth mentioning.
df.rename(columns=lambda x, *, y=iter(new): next(y))
x098 y765 z432
0 1 3 5
1 2 4 6
Column names vs Names of Series
I would like to explain a bit what happens behind the scenes.
Dataframes are a set of Series.
Series in turn are an extension of a numpy.array.
numpy.arrays have a property .name.
This is the name of the series. It is seldom that Pandas respects this attribute, but it lingers in places and can be used to hack some Pandas behaviors.
Naming the list of columns
A lot of answers here talks about the df.columns attribute being a list when in fact it is a Series. This means it has a .name attribute.
This is what happens if you decide to fill in the name of the columns Series:
df.columns = ['column_one', 'column_two']
df.columns.names = ['name of the list of columns']
df.index.names = ['name of the index']
name of the list of columns column_one column_two
name of the index
0 4 1
1 5 2
2 6 3
Note that the name of the index always comes one column lower.
Artefacts that linger
The .name attribute lingers on sometimes. If you set df.columns = ['one', 'two'] then the df.one.name will be 'one'.
If you set df.one.name = 'three' then df.columns will still give you ['one', 'two'], and df.one.name will give you 'three'.
BUT
pd.DataFrame(df.one) will return
three
0 1
1 2
2 3
Because Pandas reuses the .name of the already defined Series.
Multi-level column names
Pandas has ways of doing multi-layered column names. There is not so much magic involved, but I wanted to cover this in my answer too since I don't see anyone picking up on this here.
|one |
|one |two |
0 | 4 | 1 |
1 | 5 | 2 |
2 | 6 | 3 |
This is easily achievable by setting columns to lists, like this:
df.columns = [['one', 'one'], ['one', 'two']]
Many of pandas functions have an inplace parameter. When setting it True, the transformation applies directly to the dataframe that you are calling it on. For example:
df = pd.DataFrame({'$a':[1,2], '$b': [3,4]})
df.rename(columns={'$a': 'a'}, inplace=True)
df.columns
>>> Index(['a', '$b'], dtype='object')
Alternatively, there are cases where you want to preserve the original dataframe. I have often seen people fall into this case if creating the dataframe is an expensive task. For example, if creating the dataframe required querying a snowflake database. In this case, just make sure the the inplace parameter is set to False.
df = pd.DataFrame({'$a':[1,2], '$b': [3,4]})
df2 = df.rename(columns={'$a': 'a'}, inplace=False)
df.columns
>>> Index(['$a', '$b'], dtype='object')
df2.columns
>>> Index(['a', '$b'], dtype='object')
If these types of transformations are something that you do often, you could also look into a number of different pandas GUI tools. I'm the creator of one called Mito. It’s a spreadsheet that automatically converts your edits to python code.
Let's understand renaming by a small example...
Renaming columns using mapping:
df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6]}) # Creating a df with column name A and B
df.rename({"A": "new_a", "B": "new_b"}, axis='columns', inplace =True) # Renaming column A with 'new_a' and B with 'new_b'
Output:
new_a new_b
0 1 4
1 2 5
2 3 6
Renaming index/Row_Name using mapping:
df.rename({0: "x", 1: "y", 2: "z"}, axis='index', inplace =True) # Row name are getting replaced by 'x', 'y', and 'z'.
Output:
new_a new_b
x 1 4
y 2 5
z 3 6
Suppose your dataset name is df, and df has.
df = ['$a', '$b', '$c', '$d', '$e']`
So, to rename these, we would simply do.
df.columns = ['a','b','c','d','e']
Let's say this is your dataframe.
You can rename the columns using two methods.
Using dataframe.columns=[#list]
df.columns=['a','b','c','d','e']
The limitation of this method is that if one column has to be changed, full column list has to be passed. Also, this method is not applicable on index labels.
For example, if you passed this:
df.columns = ['a','b','c','d']
This will throw an error. Length mismatch: Expected axis has 5 elements, new values have 4 elements.
Another method is the Pandas rename() method which is used to rename any index, column or row
df = df.rename(columns={'$a':'a'})
Similarly, you can change any rows or columns.
If you've got the dataframe, df.columns dumps everything into a list you can manipulate and then reassign into your dataframe as the names of columns...
columns = df.columns
columns = [row.replace("$", "") for row in columns]
df.rename(columns=dict(zip(columns, things)), inplace=True)
df.head() # To validate the output
Best way? I don't know. A way - yes.
A better way of evaluating all the main techniques put forward in the answers to the question is below using cProfile to gage memory and execution time. #kadee, #kaitlyn, and #eumiro had the functions with the fastest execution times - though these functions are so fast we're comparing the rounding of 0.000 and 0.001 seconds for all the answers. Moral: my answer above likely isn't the 'best' way.
import pandas as pd
import cProfile, pstats, re
old_names = ['$a', '$b', '$c', '$d', '$e']
new_names = ['a', 'b', 'c', 'd', 'e']
col_dict = {'$a': 'a', '$b': 'b', '$c': 'c', '$d': 'd', '$e': 'e'}
df = pd.DataFrame({'$a':[1, 2], '$b': [10, 20], '$c': ['bleep', 'blorp'], '$d': [1, 2], '$e': ['texa$', '']})
df.head()
def eumiro(df, nn):
df.columns = nn
# This direct renaming approach is duplicated in methodology in several other answers:
return df
def lexual1(df):
return df.rename(columns=col_dict)
def lexual2(df, col_dict):
return df.rename(columns=col_dict, inplace=True)
def Panda_Master_Hayden(df):
return df.rename(columns=lambda x: x[1:], inplace=True)
def paulo1(df):
return df.rename(columns=lambda x: x.replace('$', ''))
def paulo2(df):
return df.rename(columns=lambda x: x.replace('$', ''), inplace=True)
def migloo(df, on, nn):
return df.rename(columns=dict(zip(on, nn)), inplace=True)
def kadee(df):
return df.columns.str.replace('$', '')
def awo(df):
columns = df.columns
columns = [row.replace("$", "") for row in columns]
return df.rename(columns=dict(zip(columns, '')), inplace=True)
def kaitlyn(df):
df.columns = [col.strip('$') for col in df.columns]
return df
print 'eumiro'
cProfile.run('eumiro(df, new_names)')
print 'lexual1'
cProfile.run('lexual1(df)')
print 'lexual2'
cProfile.run('lexual2(df, col_dict)')
print 'andy hayden'
cProfile.run('Panda_Master_Hayden(df)')
print 'paulo1'
cProfile.run('paulo1(df)')
print 'paulo2'
cProfile.run('paulo2(df)')
print 'migloo'
cProfile.run('migloo(df, old_names, new_names)')
print 'kadee'
cProfile.run('kadee(df)')
print 'awo'
cProfile.run('awo(df)')
print 'kaitlyn'
cProfile.run('kaitlyn(df)')
df = pd.DataFrame({'$a': [1], '$b': [1], '$c': [1], '$d': [1], '$e': [1]})
If your new list of columns is in the same order as the existing columns, the assignment is simple:
new_cols = ['a', 'b', 'c', 'd', 'e']
df.columns = new_cols
>>> df
a b c d e
0 1 1 1 1 1
If you had a dictionary keyed on old column names to new column names, you could do the following:
d = {'$a': 'a', '$b': 'b', '$c': 'c', '$d': 'd', '$e': 'e'}
df.columns = df.columns.map(lambda col: d[col]) # Or `.map(d.get)` as pointed out by #PiRSquared.
>>> df
a b c d e
0 1 1 1 1 1
If you don't have a list or dictionary mapping, you could strip the leading $ symbol via a list comprehension:
df.columns = [col[1:] if col[0] == '$' else col for col in df]
df.rename(index=str, columns={'A':'a', 'B':'b'})
pandas.DataFrame.rename
If you already have a list for the new column names, you can try this:
new_cols = ['a', 'b', 'c', 'd', 'e']
new_names_map = {df.columns[i]:new_cols[i] for i in range(len(new_cols))}
df.rename(new_names_map, axis=1, inplace=True)
Another way we could replace the original column labels is by stripping the unwanted characters (here '$') from the original column labels.
This could have been done by running a for loop over df.columns and appending the stripped columns to df.columns.
Instead, we can do this neatly in a single statement by using list comprehension like below:
df.columns = [col.strip('$') for col in df.columns]
(strip method in Python strips the given character from beginning and end of the string.)
It is real simple. Just use:
df.columns = ['Name1', 'Name2', 'Name3'...]
And it will assign the column names by the order you put them in.
# This way it will work
import pandas as pd
# Define a dictionary
rankings = {'test': ['a'],
'odi': ['E'],
't20': ['P']}
# Convert the dictionary into DataFrame
rankings_pd = pd.DataFrame(rankings)
# Before renaming the columns
print(rankings_pd)
rankings_pd.rename(columns = {'test':'TEST'}, inplace = True)
You could use str.slice for that:
df.columns = df.columns.str.slice(1)
Another option is to rename using a regular expression:
import pandas as pd
import re
df = pd.DataFrame({'$a':[1,2], '$b':[3,4], '$c':[5,6]})
df = df.rename(columns=lambda x: re.sub('\$','',x))
>>> df
a b c
0 1 3 5
1 2 4 6
My method is generic wherein you can add additional delimiters by comma separating delimiters= variable and future-proof it.
Working Code:
import pandas as pd
import re
df = pd.DataFrame({'$a':[1,2], '$b': [3,4],'$c':[5,6], '$d': [7,8], '$e': [9,10]})
delimiters = '$'
matchPattern = '|'.join(map(re.escape, delimiters))
df.columns = [re.split(matchPattern, i)[1] for i in df.columns ]
Output:
>>> df
$a $b $c $d $e
0 1 3 5 7 9
1 2 4 6 8 10
>>> df
a b c d e
0 1 3 5 7 9
1 2 4 6 8 10
Note that the approaches in previous answers do not work for a MultiIndex. For a MultiIndex, you need to do something like the following:
>>> df = pd.DataFrame({('$a','$x'):[1,2], ('$b','$y'): [3,4], ('e','f'):[5,6]})
>>> df
$a $b e
$x $y f
0 1 3 5
1 2 4 6
>>> rename = {('$a','$x'):('a','x'), ('$b','$y'):('b','y')}
>>> df.columns = pandas.MultiIndex.from_tuples([
rename.get(item, item) for item in df.columns.tolist()])
>>> df
a b e
x y f
0 1 3 5
1 2 4 6
If you have to deal with loads of columns named by the providing system out of your control, I came up with the following approach that is a combination of a general approach and specific replacements in one go.
First create a dictionary from the dataframe column names using regular expressions in order to throw away certain appendixes of column names and then add specific replacements to the dictionary to name core columns as expected later in the receiving database.
This is then applied to the dataframe in one go.
dict = dict(zip(df.columns, df.columns.str.replace('(:S$|:C1$|:L$|:D$|\.Serial:L$)', '')))
dict['brand_timeseries:C1'] = 'BTS'
dict['respid:L'] = 'RespID'
dict['country:C1'] = 'CountryID'
dict['pim1:D'] = 'pim_actual'
df.rename(columns=dict, inplace=True)
If you just want to remove the '$' sign then use the below code
df.columns = pd.Series(df.columns.str.replace("$", ""))
In addition to the solution already provided, you can replace all the columns while you are reading the file. We can use names and header=0 to do that.
First, we create a list of the names that we like to use as our column names:
import pandas as pd
ufo_cols = ['city', 'color reported', 'shape reported', 'state', 'time']
ufo.columns = ufo_cols
ufo = pd.read_csv('link to the file you are using', names = ufo_cols, header = 0)
In this case, all the column names will be replaced with the names you have in your list.
Here's a nifty little function I like to use to cut down on typing:
def rename(data, oldnames, newname):
if type(oldnames) == str: # Input can be a string or list of strings
oldnames = [oldnames] # When renaming multiple columns
newname = [newname] # Make sure you pass the corresponding list of new names
i = 0
for name in oldnames:
oldvar = [c for c in data.columns if name in c]
if len(oldvar) == 0:
raise ValueError("Sorry, couldn't find that column in the dataset")
if len(oldvar) > 1: # Doesn't have to be an exact match
print("Found multiple columns that matched " + str(name) + ": ")
for c in oldvar:
print(str(oldvar.index(c)) + ": " + str(c))
ind = input('Please enter the index of the column you would like to rename: ')
oldvar = oldvar[int(ind)]
if len(oldvar) == 1:
oldvar = oldvar[0]
data = data.rename(columns = {oldvar : newname[i]})
i += 1
return data
Here is an example of how it works:
In [2]: df = pd.DataFrame(np.random.randint(0, 10, size=(10, 4)), columns = ['col1', 'col2', 'omg', 'idk'])
# First list = existing variables
# Second list = new names for those variables
In [3]: df = rename(df, ['col', 'omg'],['first', 'ohmy'])
Found multiple columns that matched col:
0: col1
1: col2
Please enter the index of the column you would like to rename: 0
In [4]: df.columns
Out[5]: Index(['first', 'col2', 'ohmy', 'idk'], dtype='object')
How do I create an empty DataFrame, then add rows, one by one?
I created an empty DataFrame:
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
Then I can add a new row at the end and fill a single field with:
df = df._set_value(index=len(df), col='qty1', value=10.0)
It works for only one field at a time. What is a better way to add new row to df?
You can use df.loc[i], where the row with index i will be what you specify it to be in the dataframe.
>>> import pandas as pd
>>> from numpy.random import randint
>>> df = pd.DataFrame(columns=['lib', 'qty1', 'qty2'])
>>> for i in range(5):
>>> df.loc[i] = ['name' + str(i)] + list(randint(10, size=2))
>>> df
lib qty1 qty2
0 name0 3 3
1 name1 2 4
2 name2 2 8
3 name3 2 1
4 name4 9 6
In case you can get all data for the data frame upfront, there is a much faster approach than appending to a data frame:
Create a list of dictionaries in which each dictionary corresponds to an input data row.
Create a data frame from this list.
I had a similar task for which appending to a data frame row by row took 30 min, and creating a data frame from a list of dictionaries completed within seconds.
rows_list = []
for row in input_rows:
dict1 = {}
# get input row in dictionary format
# key = col_name
dict1.update(blah..)
rows_list.append(dict1)
df = pd.DataFrame(rows_list)
In the case of adding a lot of rows to dataframe, I am interested in performance. So I tried the four most popular methods and checked their speed.
Performance
Using .append (NPE's answer)
Using .loc (fred's answer)
Using .loc with preallocating (FooBar's answer)
Using dict and create DataFrame in the end (ShikharDua's answer)
Runtime results (in seconds):
Approach
1000 rows
5000 rows
10 000 rows
.append
0.69
3.39
6.78
.loc without prealloc
0.74
3.90
8.35
.loc with prealloc
0.24
2.58
8.70
dict
0.012
0.046
0.084
So I use addition through the dictionary for myself.
Code:
import pandas as pd
import numpy as np
import time
del df1, df2, df3, df4
numOfRows = 1000
# append
startTime = time.perf_counter()
df1 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows-4):
df1 = df1.append( dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']), ignore_index=True)
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df1.shape)
# .loc w/o prealloc
startTime = time.perf_counter()
df2 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows):
df2.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df2.shape)
# .loc with prealloc
df3 = pd.DataFrame(index=np.arange(0, numOfRows), columns=['A', 'B', 'C', 'D', 'E'] )
startTime = time.perf_counter()
for i in range( 1,numOfRows):
df3.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df3.shape)
# dict
startTime = time.perf_counter()
row_list = []
for i in range (0,5):
row_list.append(dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']))
for i in range( 1,numOfRows-4):
dict1 = dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E'])
row_list.append(dict1)
df4 = pd.DataFrame(row_list, columns=['A','B','C','D','E'])
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df4.shape)
P.S.: I believe my realization isn't perfect, and maybe there is some optimization that could be done.
You could use pandas.concat(). For details and examples, see Merge, join, and concatenate.
For example:
def append_row(df, row):
return pd.concat([
df,
pd.DataFrame([row], columns=row.index)]
).reset_index(drop=True)
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
new_row = pd.Series({'lib':'A', 'qty1':1, 'qty2': 2})
df = append_row(df, new_row)
NEVER grow a DataFrame!
Yes, people have already explained that you should NEVER grow a DataFrame, and that you should append your data to a list and convert it to a DataFrame once at the end. But do you understand why?
Here are the most important reasons, taken from my post here.
It is always cheaper/faster to append to a list and create a DataFrame in one go.
Lists take up less memory and are a much lighter data structure to work with, append, and remove.
dtypes are automatically inferred for your data. On the flip side, creating an empty frame of NaNs will automatically make them object, which is bad.
An index is automatically created for you, instead of you having to take care to assign the correct index to the row you are appending.
This is The Right Way™ to accumulate your data
data = []
for a, b, c in some_function_that_yields_data():
data.append([a, b, c])
df = pd.DataFrame(data, columns=['A', 'B', 'C'])
These options are horrible
append or concat inside a loop
append and concat aren't inherently bad in isolation. The
problem starts when you iteratively call them inside a loop - this
results in quadratic memory usage.
# Creates empty DataFrame and appends
df = pd.DataFrame(columns=['A', 'B', 'C'])
for a, b, c in some_function_that_yields_data():
df = df.append({'A': i, 'B': b, 'C': c}, ignore_index=True)
# This is equally bad:
# df = pd.concat(
# [df, pd.Series({'A': i, 'B': b, 'C': c})],
# ignore_index=True)
Empty DataFrame of NaNs
Never create a DataFrame of NaNs as the columns are initialized with
object (slow, un-vectorizable dtype).
# Creates DataFrame of NaNs and overwrites values.
df = pd.DataFrame(columns=['A', 'B', 'C'], index=range(5))
for a, b, c in some_function_that_yields_data():
df.loc[len(df)] = [a, b, c]
The Proof is in the Pudding
Timing these methods is the fastest way to see just how much they differ in terms of their memory and utility.
Benchmarking code for reference.
It's posts like this that remind me why I'm a part of this community. People understand the importance of teaching folks getting the right answer with the right code, not the right answer with wrong code. Now you might argue that it is not an issue to use loc or append if you're only adding a single row to your DataFrame. However, people often look to this question to add more than just one row - often the requirement is to iteratively add a row inside a loop using data that comes from a function (see related question). In that case it is important to understand that iteratively growing a DataFrame is not a good idea.
If you know the number of entries ex ante, you should preallocate the space by also providing the index (taking the data example from a different answer):
import pandas as pd
import numpy as np
# we know we're gonna have 5 rows of data
numberOfRows = 5
# create dataframe
df = pd.DataFrame(index=np.arange(0, numberOfRows), columns=('lib', 'qty1', 'qty2') )
# now fill it up row by row
for x in np.arange(0, numberOfRows):
#loc or iloc both work here since the index is natural numbers
df.loc[x] = [np.random.randint(-1,1) for n in range(3)]
In[23]: df
Out[23]:
lib qty1 qty2
0 -1 -1 -1
1 0 0 0
2 -1 0 -1
3 0 -1 0
4 -1 0 0
Speed comparison
In[30]: %timeit tryThis() # function wrapper for this answer
In[31]: %timeit tryOther() # function wrapper without index (see, for example, #fred)
1000 loops, best of 3: 1.23 ms per loop
100 loops, best of 3: 2.31 ms per loop
And - as from the comments - with a size of 6000, the speed difference becomes even larger:
Increasing the size of the array (12) and the number of rows (500) makes
the speed difference more striking: 313ms vs 2.29s
mycolumns = ['A', 'B']
df = pd.DataFrame(columns=mycolumns)
rows = [[1,2],[3,4],[5,6]]
for row in rows:
df.loc[len(df)] = row
You can append a single row as a dictionary using the ignore_index option.
>>> f = pandas.DataFrame(data = {'Animal':['cow','horse'], 'Color':['blue', 'red']})
>>> f
Animal Color
0 cow blue
1 horse red
>>> f.append({'Animal':'mouse', 'Color':'black'}, ignore_index=True)
Animal Color
0 cow blue
1 horse red
2 mouse black
For efficient appending, see How to add an extra row to a pandas dataframe and Setting With Enlargement.
Add rows through loc/ix on non existing key index data. For example:
In [1]: se = pd.Series([1,2,3])
In [2]: se
Out[2]:
0 1
1 2
2 3
dtype: int64
In [3]: se[5] = 5.
In [4]: se
Out[4]:
0 1.0
1 2.0
2 3.0
5 5.0
dtype: float64
Or:
In [1]: dfi = pd.DataFrame(np.arange(6).reshape(3,2),
.....: columns=['A','B'])
.....:
In [2]: dfi
Out[2]:
A B
0 0 1
1 2 3
2 4 5
In [3]: dfi.loc[:,'C'] = dfi.loc[:,'A']
In [4]: dfi
Out[4]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
In [5]: dfi.loc[3] = 5
In [6]: dfi
Out[6]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
3 5 5 5
For the sake of a Pythonic way:
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
res = res.append([{'qty1':10.0}], ignore_index=True)
print(res.head())
lib qty1 qty2
0 NaN 10.0 NaN
You can also build up a list of lists and convert it to a dataframe -
import pandas as pd
columns = ['i','double','square']
rows = []
for i in range(6):
row = [i, i*2, i*i]
rows.append(row)
df = pd.DataFrame(rows, columns=columns)
giving
i double square
0 0 0 0
1 1 2 1
2 2 4 4
3 3 6 9
4 4 8 16
5 5 10 25
If you always want to add a new row at the end, use this:
df.loc[len(df)] = ['name5', 9, 0]
I figured out a simple and nice way:
>>> df
A B C
one 1 2 3
>>> df.loc["two"] = [4,5,6]
>>> df
A B C
one 1 2 3
two 4 5 6
Note the caveat with performance as noted in the comments.
This is not an answer to the OP question, but a toy example to illustrate ShikharDua's answer which I found very useful.
While this fragment is trivial, in the actual data I had 1,000s of rows, and many columns, and I wished to be able to group by different columns and then perform the statistics below for more than one target column. So having a reliable method for building the data frame one row at a time was a great convenience. Thank you ShikharDua!
import pandas as pd
BaseData = pd.DataFrame({ 'Customer' : ['Acme','Mega','Acme','Acme','Mega','Acme'],
'Territory' : ['West','East','South','West','East','South'],
'Product' : ['Econ','Luxe','Econ','Std','Std','Econ']})
BaseData
columns = ['Customer','Num Unique Products', 'List Unique Products']
rows_list=[]
for name, group in BaseData.groupby('Customer'):
RecordtoAdd={} #initialise an empty dict
RecordtoAdd.update({'Customer' : name}) #
RecordtoAdd.update({'Num Unique Products' : len(pd.unique(group['Product']))})
RecordtoAdd.update({'List Unique Products' : pd.unique(group['Product'])})
rows_list.append(RecordtoAdd)
AnalysedData = pd.DataFrame(rows_list)
print('Base Data : \n',BaseData,'\n\n Analysed Data : \n',AnalysedData)
You can use a generator object to create a Dataframe, which will be more memory efficient over the list.
num = 10
# Generator function to generate generator object
def numgen_func(num):
for i in range(num):
yield ('name_{}'.format(i), (i*i), (i*i*i))
# Generator expression to generate generator object (Only once data get populated, can not be re used)
numgen_expression = (('name_{}'.format(i), (i*i), (i*i*i)) for i in range(num) )
df = pd.DataFrame(data=numgen_func(num), columns=('lib', 'qty1', 'qty2'))
To add raw to existing DataFrame you can use append method.
df = df.append([{ 'lib': "name_20", 'qty1': 20, 'qty2': 400 }])
Instead of a list of dictionaries as in ShikharDua's answer (row-based), we can also represent our table as a dictionary of lists (column-based), where each list stores one column in row-order, given we know our columns beforehand. At the end we construct our DataFrame once.
In both cases, the dictionary keys are always the column names. Row order is stored implicitly as order in a list. For c columns and n rows, this uses one dictionary of c lists, versus one list of n dictionaries. The list-of-dictionaries method has each dictionary storing all keys redundantly and requires creating a new dictionary for every row. Here we only append to lists, which overall is the same time complexity (adding entries to list and dictionary are both amortized constant time) but may have less overhead due to being a simple operation.
# Current data
data = {"Animal":["cow", "horse"], "Color":["blue", "red"]}
# Adding a new row (be careful to ensure every column gets another value)
data["Animal"].append("mouse")
data["Color"].append("black")
# At the end, construct our DataFrame
df = pd.DataFrame(data)
# Animal Color
# 0 cow blue
# 1 horse red
# 2 mouse black
Create a new record (data frame) and add to old_data_frame.
Pass a list of values and the corresponding column names to create a new_record (data_frame):
new_record = pd.DataFrame([[0, 'abcd', 0, 1, 123]], columns=['a', 'b', 'c', 'd', 'e'])
old_data_frame = pd.concat([old_data_frame, new_record])
Here is the way to add/append a row in a Pandas DataFrame:
def add_row(df, row):
df.loc[-1] = row
df.index = df.index + 1
return df.sort_index()
add_row(df, [1,2,3])
It can be used to insert/append a row in an empty or populated Pandas DataFrame.
If you want to add a row at the end, append it as a list:
valuestoappend = [va1, val2, val3]
res = res.append(pd.Series(valuestoappend, index = ['lib', 'qty1', 'qty2']), ignore_index = True)
Another way to do it (probably not very performant):
# add a row
def add_row(df, row):
colnames = list(df.columns)
ncol = len(colnames)
assert ncol == len(row), "Length of row must be the same as width of DataFrame: %s" % row
return df.append(pd.DataFrame([row], columns=colnames))
You can also enhance the DataFrame class like this:
import pandas as pd
def add_row(self, row):
self.loc[len(self.index)] = row
pd.DataFrame.add_row = add_row
All you need is loc[df.shape[0]] or loc[len(df)]
# Assuming your df has 4 columns (str, int, str, bool)
df.loc[df.shape[0]] = ['col1Value', 100, 'col3Value', False]
or
df.loc[len(df)] = ['col1Value', 100, 'col3Value', False]
You can concatenate two DataFrames for this. I basically came across this problem to add a new row to an existing DataFrame with a character index (not numeric).
So, I input the data for a new row in a duct() and index in a list.
new_dict = {put input for new row here}
new_list = [put your index here]
new_df = pd.DataFrame(data=new_dict, index=new_list)
df = pd.concat([existing_df, new_df])
initial_data = {'lib': np.array([1,2,3,4]), 'qty1': [1,2,3,4], 'qty2': [1,2,3,4]}
df = pd.DataFrame(initial_data)
df
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
val_1 = [10]
val_2 = [14]
val_3 = [20]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
You can use a for loop to iterate through values or can add arrays of values.
val_1 = [10, 11, 12, 13]
val_2 = [14, 15, 16, 17]
val_3 = [20, 21, 22, 43]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
1 11 15 21
2 12 16 22
3 13 17 43
Make it simple. By taking a list as input which will be appended as a row in the data-frame:
import pandas as pd
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
for i in range(5):
res_list = list(map(int, input().split()))
res = res.append(pd.Series(res_list, index=['lib', 'qty1', 'qty2']), ignore_index=True)
pandas.DataFrame.append
DataFrame.append(self, other, ignore_index=False, verify_integrity=False, sort=False) → 'DataFrame'
Code
df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
df.append(df2)
With ignore_index set to True:
df.append(df2, ignore_index=True)
If you have a data frame df and want to add a list new_list as a new row to df, you can simply do:
df.loc[len(df)] = new_list
If you want to add a new data frame new_df under data frame df, then you can use:
df.append(new_df)
We often see the construct df.loc[subscript] = … to assign to one DataFrame row. Mikhail_Sam posted benchmarks containing, among others, this construct as well as the method using dict and create DataFrame in the end. He found the latter to be the fastest by far.
But if we replace the df3.loc[i] = … (with preallocated DataFrame) in his code with df3.values[i] = …, the outcome changes significantly, in that that method performs similar to the one using dict. So we should more often take the use of df.values[subscript] = … into consideration. However note that .values takes a zero-based subscript, which may be different from the DataFrame.index.
Before going to add a row, we have to convert the dataframe to a dictionary. There you can see the keys as columns in the dataframe and the values of the columns are again stored in the dictionary, but there the key for every column is the index number in the dataframe.
That idea makes me to write the below code.
df2 = df.to_dict()
values = ["s_101", "hyderabad", 10, 20, 16, 13, 15, 12, 12, 13, 25, 26, 25, 27, "good", "bad"] # This is the total row that we are going to add
i = 0
for x in df.columns: # Here df.columns gives us the main dictionary key
df2[x][101] = values[i] # Here the 101 is our index number. It is also the key of the sub dictionary
i += 1
If all data in your Dataframe has the same dtype you might use a NumPy array. You can write rows directly into the predefined array and convert it to a dataframe at the end.
It seems to be even faster than converting a list of dicts.
import pandas as pd
import numpy as np
from string import ascii_uppercase
startTime = time.perf_counter()
numcols, numrows = 5, 10000
npdf = np.ones((numrows, numcols))
for row in range(numrows):
npdf[row, 0:] = np.random.randint(0, 100, (1, numcols))
df5 = pd.DataFrame(npdf, columns=list(ascii_uppercase[:numcols]))
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df5.shape)
This code snippet uses a list of dictionaries to update the data frame. It adds on to ShikharDua's and Mikhail_Sam's answers.
import pandas as pd
colour = ["red", "big", "tasty"]
fruits = ["apple", "banana", "cherry"]
dict1={}
feat_list=[]
for x in colour:
for y in fruits:
# print(x, y)
dict1 = dict([('x',x),('y',y)])
# print(f'dict 1 {dict1}')
feat_list.append(dict1)
# print(f'feat_list {feat_list}')
feat_df=pd.DataFrame(feat_list)
feat_df.to_csv('feat1.csv')
How do I create an empty DataFrame, then add rows, one by one?
I created an empty DataFrame:
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
Then I can add a new row at the end and fill a single field with:
df = df._set_value(index=len(df), col='qty1', value=10.0)
It works for only one field at a time. What is a better way to add new row to df?
You can use df.loc[i], where the row with index i will be what you specify it to be in the dataframe.
>>> import pandas as pd
>>> from numpy.random import randint
>>> df = pd.DataFrame(columns=['lib', 'qty1', 'qty2'])
>>> for i in range(5):
>>> df.loc[i] = ['name' + str(i)] + list(randint(10, size=2))
>>> df
lib qty1 qty2
0 name0 3 3
1 name1 2 4
2 name2 2 8
3 name3 2 1
4 name4 9 6
In case you can get all data for the data frame upfront, there is a much faster approach than appending to a data frame:
Create a list of dictionaries in which each dictionary corresponds to an input data row.
Create a data frame from this list.
I had a similar task for which appending to a data frame row by row took 30 min, and creating a data frame from a list of dictionaries completed within seconds.
rows_list = []
for row in input_rows:
dict1 = {}
# get input row in dictionary format
# key = col_name
dict1.update(blah..)
rows_list.append(dict1)
df = pd.DataFrame(rows_list)
In the case of adding a lot of rows to dataframe, I am interested in performance. So I tried the four most popular methods and checked their speed.
Performance
Using .append (NPE's answer)
Using .loc (fred's answer)
Using .loc with preallocating (FooBar's answer)
Using dict and create DataFrame in the end (ShikharDua's answer)
Runtime results (in seconds):
Approach
1000 rows
5000 rows
10 000 rows
.append
0.69
3.39
6.78
.loc without prealloc
0.74
3.90
8.35
.loc with prealloc
0.24
2.58
8.70
dict
0.012
0.046
0.084
So I use addition through the dictionary for myself.
Code:
import pandas as pd
import numpy as np
import time
del df1, df2, df3, df4
numOfRows = 1000
# append
startTime = time.perf_counter()
df1 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows-4):
df1 = df1.append( dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']), ignore_index=True)
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df1.shape)
# .loc w/o prealloc
startTime = time.perf_counter()
df2 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows):
df2.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df2.shape)
# .loc with prealloc
df3 = pd.DataFrame(index=np.arange(0, numOfRows), columns=['A', 'B', 'C', 'D', 'E'] )
startTime = time.perf_counter()
for i in range( 1,numOfRows):
df3.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df3.shape)
# dict
startTime = time.perf_counter()
row_list = []
for i in range (0,5):
row_list.append(dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']))
for i in range( 1,numOfRows-4):
dict1 = dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E'])
row_list.append(dict1)
df4 = pd.DataFrame(row_list, columns=['A','B','C','D','E'])
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df4.shape)
P.S.: I believe my realization isn't perfect, and maybe there is some optimization that could be done.
You could use pandas.concat(). For details and examples, see Merge, join, and concatenate.
For example:
def append_row(df, row):
return pd.concat([
df,
pd.DataFrame([row], columns=row.index)]
).reset_index(drop=True)
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
new_row = pd.Series({'lib':'A', 'qty1':1, 'qty2': 2})
df = append_row(df, new_row)
NEVER grow a DataFrame!
Yes, people have already explained that you should NEVER grow a DataFrame, and that you should append your data to a list and convert it to a DataFrame once at the end. But do you understand why?
Here are the most important reasons, taken from my post here.
It is always cheaper/faster to append to a list and create a DataFrame in one go.
Lists take up less memory and are a much lighter data structure to work with, append, and remove.
dtypes are automatically inferred for your data. On the flip side, creating an empty frame of NaNs will automatically make them object, which is bad.
An index is automatically created for you, instead of you having to take care to assign the correct index to the row you are appending.
This is The Right Way™ to accumulate your data
data = []
for a, b, c in some_function_that_yields_data():
data.append([a, b, c])
df = pd.DataFrame(data, columns=['A', 'B', 'C'])
These options are horrible
append or concat inside a loop
append and concat aren't inherently bad in isolation. The
problem starts when you iteratively call them inside a loop - this
results in quadratic memory usage.
# Creates empty DataFrame and appends
df = pd.DataFrame(columns=['A', 'B', 'C'])
for a, b, c in some_function_that_yields_data():
df = df.append({'A': i, 'B': b, 'C': c}, ignore_index=True)
# This is equally bad:
# df = pd.concat(
# [df, pd.Series({'A': i, 'B': b, 'C': c})],
# ignore_index=True)
Empty DataFrame of NaNs
Never create a DataFrame of NaNs as the columns are initialized with
object (slow, un-vectorizable dtype).
# Creates DataFrame of NaNs and overwrites values.
df = pd.DataFrame(columns=['A', 'B', 'C'], index=range(5))
for a, b, c in some_function_that_yields_data():
df.loc[len(df)] = [a, b, c]
The Proof is in the Pudding
Timing these methods is the fastest way to see just how much they differ in terms of their memory and utility.
Benchmarking code for reference.
It's posts like this that remind me why I'm a part of this community. People understand the importance of teaching folks getting the right answer with the right code, not the right answer with wrong code. Now you might argue that it is not an issue to use loc or append if you're only adding a single row to your DataFrame. However, people often look to this question to add more than just one row - often the requirement is to iteratively add a row inside a loop using data that comes from a function (see related question). In that case it is important to understand that iteratively growing a DataFrame is not a good idea.
If you know the number of entries ex ante, you should preallocate the space by also providing the index (taking the data example from a different answer):
import pandas as pd
import numpy as np
# we know we're gonna have 5 rows of data
numberOfRows = 5
# create dataframe
df = pd.DataFrame(index=np.arange(0, numberOfRows), columns=('lib', 'qty1', 'qty2') )
# now fill it up row by row
for x in np.arange(0, numberOfRows):
#loc or iloc both work here since the index is natural numbers
df.loc[x] = [np.random.randint(-1,1) for n in range(3)]
In[23]: df
Out[23]:
lib qty1 qty2
0 -1 -1 -1
1 0 0 0
2 -1 0 -1
3 0 -1 0
4 -1 0 0
Speed comparison
In[30]: %timeit tryThis() # function wrapper for this answer
In[31]: %timeit tryOther() # function wrapper without index (see, for example, #fred)
1000 loops, best of 3: 1.23 ms per loop
100 loops, best of 3: 2.31 ms per loop
And - as from the comments - with a size of 6000, the speed difference becomes even larger:
Increasing the size of the array (12) and the number of rows (500) makes
the speed difference more striking: 313ms vs 2.29s
mycolumns = ['A', 'B']
df = pd.DataFrame(columns=mycolumns)
rows = [[1,2],[3,4],[5,6]]
for row in rows:
df.loc[len(df)] = row
You can append a single row as a dictionary using the ignore_index option.
>>> f = pandas.DataFrame(data = {'Animal':['cow','horse'], 'Color':['blue', 'red']})
>>> f
Animal Color
0 cow blue
1 horse red
>>> f.append({'Animal':'mouse', 'Color':'black'}, ignore_index=True)
Animal Color
0 cow blue
1 horse red
2 mouse black
For efficient appending, see How to add an extra row to a pandas dataframe and Setting With Enlargement.
Add rows through loc/ix on non existing key index data. For example:
In [1]: se = pd.Series([1,2,3])
In [2]: se
Out[2]:
0 1
1 2
2 3
dtype: int64
In [3]: se[5] = 5.
In [4]: se
Out[4]:
0 1.0
1 2.0
2 3.0
5 5.0
dtype: float64
Or:
In [1]: dfi = pd.DataFrame(np.arange(6).reshape(3,2),
.....: columns=['A','B'])
.....:
In [2]: dfi
Out[2]:
A B
0 0 1
1 2 3
2 4 5
In [3]: dfi.loc[:,'C'] = dfi.loc[:,'A']
In [4]: dfi
Out[4]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
In [5]: dfi.loc[3] = 5
In [6]: dfi
Out[6]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
3 5 5 5
For the sake of a Pythonic way:
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
res = res.append([{'qty1':10.0}], ignore_index=True)
print(res.head())
lib qty1 qty2
0 NaN 10.0 NaN
You can also build up a list of lists and convert it to a dataframe -
import pandas as pd
columns = ['i','double','square']
rows = []
for i in range(6):
row = [i, i*2, i*i]
rows.append(row)
df = pd.DataFrame(rows, columns=columns)
giving
i double square
0 0 0 0
1 1 2 1
2 2 4 4
3 3 6 9
4 4 8 16
5 5 10 25
If you always want to add a new row at the end, use this:
df.loc[len(df)] = ['name5', 9, 0]
I figured out a simple and nice way:
>>> df
A B C
one 1 2 3
>>> df.loc["two"] = [4,5,6]
>>> df
A B C
one 1 2 3
two 4 5 6
Note the caveat with performance as noted in the comments.
This is not an answer to the OP question, but a toy example to illustrate ShikharDua's answer which I found very useful.
While this fragment is trivial, in the actual data I had 1,000s of rows, and many columns, and I wished to be able to group by different columns and then perform the statistics below for more than one target column. So having a reliable method for building the data frame one row at a time was a great convenience. Thank you ShikharDua!
import pandas as pd
BaseData = pd.DataFrame({ 'Customer' : ['Acme','Mega','Acme','Acme','Mega','Acme'],
'Territory' : ['West','East','South','West','East','South'],
'Product' : ['Econ','Luxe','Econ','Std','Std','Econ']})
BaseData
columns = ['Customer','Num Unique Products', 'List Unique Products']
rows_list=[]
for name, group in BaseData.groupby('Customer'):
RecordtoAdd={} #initialise an empty dict
RecordtoAdd.update({'Customer' : name}) #
RecordtoAdd.update({'Num Unique Products' : len(pd.unique(group['Product']))})
RecordtoAdd.update({'List Unique Products' : pd.unique(group['Product'])})
rows_list.append(RecordtoAdd)
AnalysedData = pd.DataFrame(rows_list)
print('Base Data : \n',BaseData,'\n\n Analysed Data : \n',AnalysedData)
You can use a generator object to create a Dataframe, which will be more memory efficient over the list.
num = 10
# Generator function to generate generator object
def numgen_func(num):
for i in range(num):
yield ('name_{}'.format(i), (i*i), (i*i*i))
# Generator expression to generate generator object (Only once data get populated, can not be re used)
numgen_expression = (('name_{}'.format(i), (i*i), (i*i*i)) for i in range(num) )
df = pd.DataFrame(data=numgen_func(num), columns=('lib', 'qty1', 'qty2'))
To add raw to existing DataFrame you can use append method.
df = df.append([{ 'lib': "name_20", 'qty1': 20, 'qty2': 400 }])
Instead of a list of dictionaries as in ShikharDua's answer (row-based), we can also represent our table as a dictionary of lists (column-based), where each list stores one column in row-order, given we know our columns beforehand. At the end we construct our DataFrame once.
In both cases, the dictionary keys are always the column names. Row order is stored implicitly as order in a list. For c columns and n rows, this uses one dictionary of c lists, versus one list of n dictionaries. The list-of-dictionaries method has each dictionary storing all keys redundantly and requires creating a new dictionary for every row. Here we only append to lists, which overall is the same time complexity (adding entries to list and dictionary are both amortized constant time) but may have less overhead due to being a simple operation.
# Current data
data = {"Animal":["cow", "horse"], "Color":["blue", "red"]}
# Adding a new row (be careful to ensure every column gets another value)
data["Animal"].append("mouse")
data["Color"].append("black")
# At the end, construct our DataFrame
df = pd.DataFrame(data)
# Animal Color
# 0 cow blue
# 1 horse red
# 2 mouse black
Create a new record (data frame) and add to old_data_frame.
Pass a list of values and the corresponding column names to create a new_record (data_frame):
new_record = pd.DataFrame([[0, 'abcd', 0, 1, 123]], columns=['a', 'b', 'c', 'd', 'e'])
old_data_frame = pd.concat([old_data_frame, new_record])
Here is the way to add/append a row in a Pandas DataFrame:
def add_row(df, row):
df.loc[-1] = row
df.index = df.index + 1
return df.sort_index()
add_row(df, [1,2,3])
It can be used to insert/append a row in an empty or populated Pandas DataFrame.
If you want to add a row at the end, append it as a list:
valuestoappend = [va1, val2, val3]
res = res.append(pd.Series(valuestoappend, index = ['lib', 'qty1', 'qty2']), ignore_index = True)
Another way to do it (probably not very performant):
# add a row
def add_row(df, row):
colnames = list(df.columns)
ncol = len(colnames)
assert ncol == len(row), "Length of row must be the same as width of DataFrame: %s" % row
return df.append(pd.DataFrame([row], columns=colnames))
You can also enhance the DataFrame class like this:
import pandas as pd
def add_row(self, row):
self.loc[len(self.index)] = row
pd.DataFrame.add_row = add_row
All you need is loc[df.shape[0]] or loc[len(df)]
# Assuming your df has 4 columns (str, int, str, bool)
df.loc[df.shape[0]] = ['col1Value', 100, 'col3Value', False]
or
df.loc[len(df)] = ['col1Value', 100, 'col3Value', False]
You can concatenate two DataFrames for this. I basically came across this problem to add a new row to an existing DataFrame with a character index (not numeric).
So, I input the data for a new row in a duct() and index in a list.
new_dict = {put input for new row here}
new_list = [put your index here]
new_df = pd.DataFrame(data=new_dict, index=new_list)
df = pd.concat([existing_df, new_df])
initial_data = {'lib': np.array([1,2,3,4]), 'qty1': [1,2,3,4], 'qty2': [1,2,3,4]}
df = pd.DataFrame(initial_data)
df
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
val_1 = [10]
val_2 = [14]
val_3 = [20]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
You can use a for loop to iterate through values or can add arrays of values.
val_1 = [10, 11, 12, 13]
val_2 = [14, 15, 16, 17]
val_3 = [20, 21, 22, 43]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
1 11 15 21
2 12 16 22
3 13 17 43
Make it simple. By taking a list as input which will be appended as a row in the data-frame:
import pandas as pd
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
for i in range(5):
res_list = list(map(int, input().split()))
res = res.append(pd.Series(res_list, index=['lib', 'qty1', 'qty2']), ignore_index=True)
pandas.DataFrame.append
DataFrame.append(self, other, ignore_index=False, verify_integrity=False, sort=False) → 'DataFrame'
Code
df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
df.append(df2)
With ignore_index set to True:
df.append(df2, ignore_index=True)
If you have a data frame df and want to add a list new_list as a new row to df, you can simply do:
df.loc[len(df)] = new_list
If you want to add a new data frame new_df under data frame df, then you can use:
df.append(new_df)
We often see the construct df.loc[subscript] = … to assign to one DataFrame row. Mikhail_Sam posted benchmarks containing, among others, this construct as well as the method using dict and create DataFrame in the end. He found the latter to be the fastest by far.
But if we replace the df3.loc[i] = … (with preallocated DataFrame) in his code with df3.values[i] = …, the outcome changes significantly, in that that method performs similar to the one using dict. So we should more often take the use of df.values[subscript] = … into consideration. However note that .values takes a zero-based subscript, which may be different from the DataFrame.index.
Before going to add a row, we have to convert the dataframe to a dictionary. There you can see the keys as columns in the dataframe and the values of the columns are again stored in the dictionary, but there the key for every column is the index number in the dataframe.
That idea makes me to write the below code.
df2 = df.to_dict()
values = ["s_101", "hyderabad", 10, 20, 16, 13, 15, 12, 12, 13, 25, 26, 25, 27, "good", "bad"] # This is the total row that we are going to add
i = 0
for x in df.columns: # Here df.columns gives us the main dictionary key
df2[x][101] = values[i] # Here the 101 is our index number. It is also the key of the sub dictionary
i += 1
If all data in your Dataframe has the same dtype you might use a NumPy array. You can write rows directly into the predefined array and convert it to a dataframe at the end.
It seems to be even faster than converting a list of dicts.
import pandas as pd
import numpy as np
from string import ascii_uppercase
startTime = time.perf_counter()
numcols, numrows = 5, 10000
npdf = np.ones((numrows, numcols))
for row in range(numrows):
npdf[row, 0:] = np.random.randint(0, 100, (1, numcols))
df5 = pd.DataFrame(npdf, columns=list(ascii_uppercase[:numcols]))
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df5.shape)
This code snippet uses a list of dictionaries to update the data frame. It adds on to ShikharDua's and Mikhail_Sam's answers.
import pandas as pd
colour = ["red", "big", "tasty"]
fruits = ["apple", "banana", "cherry"]
dict1={}
feat_list=[]
for x in colour:
for y in fruits:
# print(x, y)
dict1 = dict([('x',x),('y',y)])
# print(f'dict 1 {dict1}')
feat_list.append(dict1)
# print(f'feat_list {feat_list}')
feat_df=pd.DataFrame(feat_list)
feat_df.to_csv('feat1.csv')
How do I create an empty DataFrame, then add rows, one by one?
I created an empty DataFrame:
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
Then I can add a new row at the end and fill a single field with:
df = df._set_value(index=len(df), col='qty1', value=10.0)
It works for only one field at a time. What is a better way to add new row to df?
You can use df.loc[i], where the row with index i will be what you specify it to be in the dataframe.
>>> import pandas as pd
>>> from numpy.random import randint
>>> df = pd.DataFrame(columns=['lib', 'qty1', 'qty2'])
>>> for i in range(5):
>>> df.loc[i] = ['name' + str(i)] + list(randint(10, size=2))
>>> df
lib qty1 qty2
0 name0 3 3
1 name1 2 4
2 name2 2 8
3 name3 2 1
4 name4 9 6
In case you can get all data for the data frame upfront, there is a much faster approach than appending to a data frame:
Create a list of dictionaries in which each dictionary corresponds to an input data row.
Create a data frame from this list.
I had a similar task for which appending to a data frame row by row took 30 min, and creating a data frame from a list of dictionaries completed within seconds.
rows_list = []
for row in input_rows:
dict1 = {}
# get input row in dictionary format
# key = col_name
dict1.update(blah..)
rows_list.append(dict1)
df = pd.DataFrame(rows_list)
In the case of adding a lot of rows to dataframe, I am interested in performance. So I tried the four most popular methods and checked their speed.
Performance
Using .append (NPE's answer)
Using .loc (fred's answer)
Using .loc with preallocating (FooBar's answer)
Using dict and create DataFrame in the end (ShikharDua's answer)
Runtime results (in seconds):
Approach
1000 rows
5000 rows
10 000 rows
.append
0.69
3.39
6.78
.loc without prealloc
0.74
3.90
8.35
.loc with prealloc
0.24
2.58
8.70
dict
0.012
0.046
0.084
So I use addition through the dictionary for myself.
Code:
import pandas as pd
import numpy as np
import time
del df1, df2, df3, df4
numOfRows = 1000
# append
startTime = time.perf_counter()
df1 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows-4):
df1 = df1.append( dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']), ignore_index=True)
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df1.shape)
# .loc w/o prealloc
startTime = time.perf_counter()
df2 = pd.DataFrame(np.random.randint(100, size=(5,5)), columns=['A', 'B', 'C', 'D', 'E'])
for i in range( 1,numOfRows):
df2.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df2.shape)
# .loc with prealloc
df3 = pd.DataFrame(index=np.arange(0, numOfRows), columns=['A', 'B', 'C', 'D', 'E'] )
startTime = time.perf_counter()
for i in range( 1,numOfRows):
df3.loc[i] = np.random.randint(100, size=(1,5))[0]
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df3.shape)
# dict
startTime = time.perf_counter()
row_list = []
for i in range (0,5):
row_list.append(dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E']))
for i in range( 1,numOfRows-4):
dict1 = dict( (a,np.random.randint(100)) for a in ['A','B','C','D','E'])
row_list.append(dict1)
df4 = pd.DataFrame(row_list, columns=['A','B','C','D','E'])
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df4.shape)
P.S.: I believe my realization isn't perfect, and maybe there is some optimization that could be done.
You could use pandas.concat(). For details and examples, see Merge, join, and concatenate.
For example:
def append_row(df, row):
return pd.concat([
df,
pd.DataFrame([row], columns=row.index)]
).reset_index(drop=True)
df = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
new_row = pd.Series({'lib':'A', 'qty1':1, 'qty2': 2})
df = append_row(df, new_row)
NEVER grow a DataFrame!
Yes, people have already explained that you should NEVER grow a DataFrame, and that you should append your data to a list and convert it to a DataFrame once at the end. But do you understand why?
Here are the most important reasons, taken from my post here.
It is always cheaper/faster to append to a list and create a DataFrame in one go.
Lists take up less memory and are a much lighter data structure to work with, append, and remove.
dtypes are automatically inferred for your data. On the flip side, creating an empty frame of NaNs will automatically make them object, which is bad.
An index is automatically created for you, instead of you having to take care to assign the correct index to the row you are appending.
This is The Right Way™ to accumulate your data
data = []
for a, b, c in some_function_that_yields_data():
data.append([a, b, c])
df = pd.DataFrame(data, columns=['A', 'B', 'C'])
These options are horrible
append or concat inside a loop
append and concat aren't inherently bad in isolation. The
problem starts when you iteratively call them inside a loop - this
results in quadratic memory usage.
# Creates empty DataFrame and appends
df = pd.DataFrame(columns=['A', 'B', 'C'])
for a, b, c in some_function_that_yields_data():
df = df.append({'A': i, 'B': b, 'C': c}, ignore_index=True)
# This is equally bad:
# df = pd.concat(
# [df, pd.Series({'A': i, 'B': b, 'C': c})],
# ignore_index=True)
Empty DataFrame of NaNs
Never create a DataFrame of NaNs as the columns are initialized with
object (slow, un-vectorizable dtype).
# Creates DataFrame of NaNs and overwrites values.
df = pd.DataFrame(columns=['A', 'B', 'C'], index=range(5))
for a, b, c in some_function_that_yields_data():
df.loc[len(df)] = [a, b, c]
The Proof is in the Pudding
Timing these methods is the fastest way to see just how much they differ in terms of their memory and utility.
Benchmarking code for reference.
It's posts like this that remind me why I'm a part of this community. People understand the importance of teaching folks getting the right answer with the right code, not the right answer with wrong code. Now you might argue that it is not an issue to use loc or append if you're only adding a single row to your DataFrame. However, people often look to this question to add more than just one row - often the requirement is to iteratively add a row inside a loop using data that comes from a function (see related question). In that case it is important to understand that iteratively growing a DataFrame is not a good idea.
If you know the number of entries ex ante, you should preallocate the space by also providing the index (taking the data example from a different answer):
import pandas as pd
import numpy as np
# we know we're gonna have 5 rows of data
numberOfRows = 5
# create dataframe
df = pd.DataFrame(index=np.arange(0, numberOfRows), columns=('lib', 'qty1', 'qty2') )
# now fill it up row by row
for x in np.arange(0, numberOfRows):
#loc or iloc both work here since the index is natural numbers
df.loc[x] = [np.random.randint(-1,1) for n in range(3)]
In[23]: df
Out[23]:
lib qty1 qty2
0 -1 -1 -1
1 0 0 0
2 -1 0 -1
3 0 -1 0
4 -1 0 0
Speed comparison
In[30]: %timeit tryThis() # function wrapper for this answer
In[31]: %timeit tryOther() # function wrapper without index (see, for example, #fred)
1000 loops, best of 3: 1.23 ms per loop
100 loops, best of 3: 2.31 ms per loop
And - as from the comments - with a size of 6000, the speed difference becomes even larger:
Increasing the size of the array (12) and the number of rows (500) makes
the speed difference more striking: 313ms vs 2.29s
mycolumns = ['A', 'B']
df = pd.DataFrame(columns=mycolumns)
rows = [[1,2],[3,4],[5,6]]
for row in rows:
df.loc[len(df)] = row
You can append a single row as a dictionary using the ignore_index option.
>>> f = pandas.DataFrame(data = {'Animal':['cow','horse'], 'Color':['blue', 'red']})
>>> f
Animal Color
0 cow blue
1 horse red
>>> f.append({'Animal':'mouse', 'Color':'black'}, ignore_index=True)
Animal Color
0 cow blue
1 horse red
2 mouse black
For efficient appending, see How to add an extra row to a pandas dataframe and Setting With Enlargement.
Add rows through loc/ix on non existing key index data. For example:
In [1]: se = pd.Series([1,2,3])
In [2]: se
Out[2]:
0 1
1 2
2 3
dtype: int64
In [3]: se[5] = 5.
In [4]: se
Out[4]:
0 1.0
1 2.0
2 3.0
5 5.0
dtype: float64
Or:
In [1]: dfi = pd.DataFrame(np.arange(6).reshape(3,2),
.....: columns=['A','B'])
.....:
In [2]: dfi
Out[2]:
A B
0 0 1
1 2 3
2 4 5
In [3]: dfi.loc[:,'C'] = dfi.loc[:,'A']
In [4]: dfi
Out[4]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
In [5]: dfi.loc[3] = 5
In [6]: dfi
Out[6]:
A B C
0 0 1 0
1 2 3 2
2 4 5 4
3 5 5 5
For the sake of a Pythonic way:
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
res = res.append([{'qty1':10.0}], ignore_index=True)
print(res.head())
lib qty1 qty2
0 NaN 10.0 NaN
You can also build up a list of lists and convert it to a dataframe -
import pandas as pd
columns = ['i','double','square']
rows = []
for i in range(6):
row = [i, i*2, i*i]
rows.append(row)
df = pd.DataFrame(rows, columns=columns)
giving
i double square
0 0 0 0
1 1 2 1
2 2 4 4
3 3 6 9
4 4 8 16
5 5 10 25
If you always want to add a new row at the end, use this:
df.loc[len(df)] = ['name5', 9, 0]
I figured out a simple and nice way:
>>> df
A B C
one 1 2 3
>>> df.loc["two"] = [4,5,6]
>>> df
A B C
one 1 2 3
two 4 5 6
Note the caveat with performance as noted in the comments.
This is not an answer to the OP question, but a toy example to illustrate ShikharDua's answer which I found very useful.
While this fragment is trivial, in the actual data I had 1,000s of rows, and many columns, and I wished to be able to group by different columns and then perform the statistics below for more than one target column. So having a reliable method for building the data frame one row at a time was a great convenience. Thank you ShikharDua!
import pandas as pd
BaseData = pd.DataFrame({ 'Customer' : ['Acme','Mega','Acme','Acme','Mega','Acme'],
'Territory' : ['West','East','South','West','East','South'],
'Product' : ['Econ','Luxe','Econ','Std','Std','Econ']})
BaseData
columns = ['Customer','Num Unique Products', 'List Unique Products']
rows_list=[]
for name, group in BaseData.groupby('Customer'):
RecordtoAdd={} #initialise an empty dict
RecordtoAdd.update({'Customer' : name}) #
RecordtoAdd.update({'Num Unique Products' : len(pd.unique(group['Product']))})
RecordtoAdd.update({'List Unique Products' : pd.unique(group['Product'])})
rows_list.append(RecordtoAdd)
AnalysedData = pd.DataFrame(rows_list)
print('Base Data : \n',BaseData,'\n\n Analysed Data : \n',AnalysedData)
You can use a generator object to create a Dataframe, which will be more memory efficient over the list.
num = 10
# Generator function to generate generator object
def numgen_func(num):
for i in range(num):
yield ('name_{}'.format(i), (i*i), (i*i*i))
# Generator expression to generate generator object (Only once data get populated, can not be re used)
numgen_expression = (('name_{}'.format(i), (i*i), (i*i*i)) for i in range(num) )
df = pd.DataFrame(data=numgen_func(num), columns=('lib', 'qty1', 'qty2'))
To add raw to existing DataFrame you can use append method.
df = df.append([{ 'lib': "name_20", 'qty1': 20, 'qty2': 400 }])
Instead of a list of dictionaries as in ShikharDua's answer (row-based), we can also represent our table as a dictionary of lists (column-based), where each list stores one column in row-order, given we know our columns beforehand. At the end we construct our DataFrame once.
In both cases, the dictionary keys are always the column names. Row order is stored implicitly as order in a list. For c columns and n rows, this uses one dictionary of c lists, versus one list of n dictionaries. The list-of-dictionaries method has each dictionary storing all keys redundantly and requires creating a new dictionary for every row. Here we only append to lists, which overall is the same time complexity (adding entries to list and dictionary are both amortized constant time) but may have less overhead due to being a simple operation.
# Current data
data = {"Animal":["cow", "horse"], "Color":["blue", "red"]}
# Adding a new row (be careful to ensure every column gets another value)
data["Animal"].append("mouse")
data["Color"].append("black")
# At the end, construct our DataFrame
df = pd.DataFrame(data)
# Animal Color
# 0 cow blue
# 1 horse red
# 2 mouse black
Create a new record (data frame) and add to old_data_frame.
Pass a list of values and the corresponding column names to create a new_record (data_frame):
new_record = pd.DataFrame([[0, 'abcd', 0, 1, 123]], columns=['a', 'b', 'c', 'd', 'e'])
old_data_frame = pd.concat([old_data_frame, new_record])
Here is the way to add/append a row in a Pandas DataFrame:
def add_row(df, row):
df.loc[-1] = row
df.index = df.index + 1
return df.sort_index()
add_row(df, [1,2,3])
It can be used to insert/append a row in an empty or populated Pandas DataFrame.
If you want to add a row at the end, append it as a list:
valuestoappend = [va1, val2, val3]
res = res.append(pd.Series(valuestoappend, index = ['lib', 'qty1', 'qty2']), ignore_index = True)
Another way to do it (probably not very performant):
# add a row
def add_row(df, row):
colnames = list(df.columns)
ncol = len(colnames)
assert ncol == len(row), "Length of row must be the same as width of DataFrame: %s" % row
return df.append(pd.DataFrame([row], columns=colnames))
You can also enhance the DataFrame class like this:
import pandas as pd
def add_row(self, row):
self.loc[len(self.index)] = row
pd.DataFrame.add_row = add_row
All you need is loc[df.shape[0]] or loc[len(df)]
# Assuming your df has 4 columns (str, int, str, bool)
df.loc[df.shape[0]] = ['col1Value', 100, 'col3Value', False]
or
df.loc[len(df)] = ['col1Value', 100, 'col3Value', False]
You can concatenate two DataFrames for this. I basically came across this problem to add a new row to an existing DataFrame with a character index (not numeric).
So, I input the data for a new row in a duct() and index in a list.
new_dict = {put input for new row here}
new_list = [put your index here]
new_df = pd.DataFrame(data=new_dict, index=new_list)
df = pd.concat([existing_df, new_df])
initial_data = {'lib': np.array([1,2,3,4]), 'qty1': [1,2,3,4], 'qty2': [1,2,3,4]}
df = pd.DataFrame(initial_data)
df
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
val_1 = [10]
val_2 = [14]
val_3 = [20]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
You can use a for loop to iterate through values or can add arrays of values.
val_1 = [10, 11, 12, 13]
val_2 = [14, 15, 16, 17]
val_3 = [20, 21, 22, 43]
df.append(pd.DataFrame({'lib': val_1, 'qty1': val_2, 'qty2': val_3}))
lib qty1 qty2
0 1 1 1
1 2 2 2
2 3 3 3
3 4 4 4
0 10 14 20
1 11 15 21
2 12 16 22
3 13 17 43
Make it simple. By taking a list as input which will be appended as a row in the data-frame:
import pandas as pd
res = pd.DataFrame(columns=('lib', 'qty1', 'qty2'))
for i in range(5):
res_list = list(map(int, input().split()))
res = res.append(pd.Series(res_list, index=['lib', 'qty1', 'qty2']), ignore_index=True)
pandas.DataFrame.append
DataFrame.append(self, other, ignore_index=False, verify_integrity=False, sort=False) → 'DataFrame'
Code
df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'))
df2 = pd.DataFrame([[5, 6], [7, 8]], columns=list('AB'))
df.append(df2)
With ignore_index set to True:
df.append(df2, ignore_index=True)
If you have a data frame df and want to add a list new_list as a new row to df, you can simply do:
df.loc[len(df)] = new_list
If you want to add a new data frame new_df under data frame df, then you can use:
df.append(new_df)
We often see the construct df.loc[subscript] = … to assign to one DataFrame row. Mikhail_Sam posted benchmarks containing, among others, this construct as well as the method using dict and create DataFrame in the end. He found the latter to be the fastest by far.
But if we replace the df3.loc[i] = … (with preallocated DataFrame) in his code with df3.values[i] = …, the outcome changes significantly, in that that method performs similar to the one using dict. So we should more often take the use of df.values[subscript] = … into consideration. However note that .values takes a zero-based subscript, which may be different from the DataFrame.index.
Before going to add a row, we have to convert the dataframe to a dictionary. There you can see the keys as columns in the dataframe and the values of the columns are again stored in the dictionary, but there the key for every column is the index number in the dataframe.
That idea makes me to write the below code.
df2 = df.to_dict()
values = ["s_101", "hyderabad", 10, 20, 16, 13, 15, 12, 12, 13, 25, 26, 25, 27, "good", "bad"] # This is the total row that we are going to add
i = 0
for x in df.columns: # Here df.columns gives us the main dictionary key
df2[x][101] = values[i] # Here the 101 is our index number. It is also the key of the sub dictionary
i += 1
If all data in your Dataframe has the same dtype you might use a NumPy array. You can write rows directly into the predefined array and convert it to a dataframe at the end.
It seems to be even faster than converting a list of dicts.
import pandas as pd
import numpy as np
from string import ascii_uppercase
startTime = time.perf_counter()
numcols, numrows = 5, 10000
npdf = np.ones((numrows, numcols))
for row in range(numrows):
npdf[row, 0:] = np.random.randint(0, 100, (1, numcols))
df5 = pd.DataFrame(npdf, columns=list(ascii_uppercase[:numcols]))
print('Elapsed time: {:6.3f} seconds for {:d} rows'.format(time.perf_counter() - startTime, numOfRows))
print(df5.shape)
This code snippet uses a list of dictionaries to update the data frame. It adds on to ShikharDua's and Mikhail_Sam's answers.
import pandas as pd
colour = ["red", "big", "tasty"]
fruits = ["apple", "banana", "cherry"]
dict1={}
feat_list=[]
for x in colour:
for y in fruits:
# print(x, y)
dict1 = dict([('x',x),('y',y)])
# print(f'dict 1 {dict1}')
feat_list.append(dict1)
# print(f'feat_list {feat_list}')
feat_df=pd.DataFrame(feat_list)
feat_df.to_csv('feat1.csv')