Related
So I have a dataframe that looks something like this:
df1 = pd.DataFrame([[1,2, 3], [5,7,8], [2,5,4]])
0 1 2
0 1 2 3
1 5 7 8
2 2 5 4
I then have a function that adds 5 to a number called add5. I'm trying to create a new column in df1 that adds 5 to all the numbers in column 2 that are greater than 3. I want to use vectorization not apply as this concept is going to be expanded to a dataset with hundreds of thousands of entries and speed will be important. I can do it without the greater than 3 constraint like this:
df1['3'] = add5(df1[2])
But my goal is to do something like this:
df1['3'] = add5(df1[2]) if df1[2] > 3
Hoping someone can point me in the right direction on this. Thanks!
With Pandas, a function applied explicitly to each row typically cannot be vectorised. Even implicit loops such as pd.Series.apply will likely be inefficient. Instead, you should use true vectorised operations, which lean heavily on NumPy in both functionality and syntax.
In this case, you can use numpy.where:
df1[3] = np.where(df1[2] > 3, df1[2] + 5, df1[2])
Alternatively, you can use pd.DataFrame.loc in a couple of steps:
df1[3] = df1[2]
df1.loc[df1[2] > 3, 3] = df1[2] + 5
In each case, the term df1[2] > 3 creates a Boolean series, which is then used to mask another series.
Result:
print(df1)
0 1 2 3
0 1 2 3 3
1 5 7 8 13
2 2 5 4 9
So I have two pandas dataframes, A and B.
A is 1000 rows x 500 columns, filled with binary values indicating either presence or absence.
B is 1024 rows x 10 columns, and is a full iteration of 0's and 1's, hence having 1024 rows.
I am trying to find which rows in A, at a particular 10 columns of A, correspond with a given row in B. I need the whole row to match up, rather than element by element.
For example, I would want
A[(A.ix[:,(1,2,3,4,5,6,7,8,9,10)==(1,0,1,0,1,0,0,1,0,0)).all(axis=1)]
To return something that rows (3,5,8,11,15) in A match up with that (1,0,1,0,1,0,0,1,0,0) row of B at those particular columns (1,2,3,4,5,6,7,8,9,10)
And I want to do this over every row in B.
The best way I could figure out to do this was:
import numpy as np
for i in B:
B_array = np.array(i)
Matching_Rows = A[(A.ix[:,(1,2,3,4,5,6,7,8,9,10)] == B_array).all(axis=1)]
Matching_Rows_Index = Matching_Rows.index
This isn't terrible for one instance, but I use it in a while loop that runs around 20,000 times; therefore, it slows it down quite a bit.
I have been messing around with DataFrame.apply to no avail. Could map work better?
I was just hoping someone saw something obviously more efficient as I am fairly new to python.
Thanks and best regards!
We can abuse the fact that both dataframes have binary values 0 or 1 by collapsing the relevant columns from A and all columns from B into 1D arrays each, when considering each row as a sequence of binary numbers that could be converted to decimal number equivalents. This should reduce the problem set considerably, which would help with performance. Now, after getting those 1D arrays, we can use np.in1d to look for matches from B in A and finally np.where on it to get the matching indices.
Thus, we would have an implementation like so -
# Setup 1D arrays corresponding to selected cols from A and entire B
S = 2**np.arange(10)
A_ID = np.dot(A[range(1,11)],S)
B_ID = np.dot(B,S)
# Look for matches that exist from B_ID in A_ID, whose indices
# would be desired row indices that have matched from B
out_row_idx = np.where(np.in1d(A_ID,B_ID))[0]
Sample run -
In [157]: # Setup dataframes A and B with rows 0, 4 in A having matches from B
...: A_arr = np.random.randint(0,2,(10,14))
...: B_arr = np.random.randint(0,2,(7,10))
...:
...: B_arr[2] = A_arr[4,1:11]
...: B_arr[4] = A_arr[4,1:11]
...: B_arr[5] = A_arr[0,1:11]
...:
...: A = pd.DataFrame(A_arr)
...: B = pd.DataFrame(B_arr)
...:
In [158]: S = 2**np.arange(10)
...: A_ID = np.dot(A[range(1,11)],S)
...: B_ID = np.dot(B,S)
...: out_row_idx = np.where(np.in1d(A_ID,B_ID))[0]
...:
In [159]: out_row_idx
Out[159]: array([0, 4])
You can use merge with reset_index - output are indexes of B which are equal in A in custom columns:
A = pd.DataFrame({'A':[1,0,1,1],
'B':[0,0,1,1],
'C':[1,0,1,1],
'D':[1,1,1,0],
'E':[1,1,0,1]})
print (A)
A B C D E
0 1 0 1 1 1
1 0 0 0 1 1
2 1 1 1 1 0
3 1 1 1 0 1
B = pd.DataFrame({'0':[1,0,1],
'1':[1,0,1],
'2':[1,0,0]})
print (B)
0 1 2
0 1 1 1
1 0 0 0
2 1 1 0
print (pd.merge(B.reset_index(),
A.reset_index(),
left_on=B.columns.tolist(),
right_on=A.columns[[0,1,2]].tolist(),
suffixes=('_B','_A')))
index_B 0 1 2 index_A A B C D E
0 0 1 1 1 2 1 1 1 1 0
1 0 1 1 1 3 1 1 1 0 1
2 1 0 0 0 1 0 0 0 1 1
print (pd.merge(B.reset_index(),
A.reset_index(),
left_on=B.columns.tolist(),
right_on=A.columns[[0,1,2]].tolist(),
suffixes=('_B','_A'))[['index_B','index_A']])
index_B index_A
0 0 2
1 0 3
2 1 1
You can do it in pandas by using loc or ix and telling it to find the rows where the ten columns are all equal. Like this:
A.loc[(A[1]==B[1]) & (A[2]==B[2]) & (A[3]==B[3]) & A[4]==B[4]) & (A[5]==B[5]) & (A[6]==B[6]) & (A[7]==B[7]) & (A[8]==B[8]) & (A[9]==B[9]) & (A[10]==B[10])]
This is quite ugly in my opinion but it will work and gets rid of the loop so it should be significantly faster. I wouldn't be surprised if someone could come up with a more elegant way of coding the same operation.
In this special case, your rows of 10 zeros and ones can be interpreted as 10 digit binaries. If B is in order, then it can be interpreted as a range from 0 to 1023. In this case, all we need to do is take A's rows in 10 column chunks and calculate what its binary equivalent is.
I'll start by defining a range of powers of two so I can do matrix multiplication with it.
twos = pd.Series(np.power(2, np.arange(10)))
Next, I'll relabel A's columns into a MultiIndex and stack to get my chunks of 10.
A = pd.DataFrame(np.random.binomial(1, .5, (1000, 500)))
A.columns = pd.MultiIndex.from_tuples(zip((A.columns / 10).tolist(), (A.columns % 10).tolist()))
A_ = A.stack(0)
A_.head()
Finally, I'll multiply A_ with twos to get integer representation of each row and unstack.
A_.dot(twos).unstack()
This is now a 1000 x 50 DataFrame where each cell represents which of B's rows we matched for that particular 10 column chunk for that particular row of A. There isn't even a need for B.
I am iterating through the rows of a pandas DataFrame, expanding each one out into N rows with additional info on each one (for simplicity I've made it a random number here):
from pandas import DataFrame
import pandas as pd
from numpy import random, arange
N=3
x = DataFrame.from_dict({'farm' : ['A','B','A','B'],
'fruit':['apple','apple','pear','pear']})
out = DataFrame()
for i,row in x.iterrows():
rows = pd.concat([row]*N).reset_index(drop=True) # requires row to be a DataFrame
out = out.append(rows.join(DataFrame({'iter': arange(N), 'value': random.uniform(size=N)})))
In this loop, row is a Series object, so the call to pd.concat doesn't work. How do I convert it to a DataFrame? (Eg. the difference between x.ix[0:0] and x.ix[0])
Thanks!
Given what you commented, I would try
def giveMeSomeRows(group):
return random.uniform(low=group.low, high=group.high, size=N)
results = x.groupby(['farm', 'fruit']).apply(giveMeSomeRows)
This should give you a separate result dataframe. I have assumed that every farm-fruit combination is unique... there might be other ways, if we'd know more about your data.
Update
Running code example
def giveMeSomeRows(group):
return random.uniform(low=group.low, high=group.high, size=N)
N = 3
df = pd.DataFrame(arange(0,8).reshape(4,2), columns=['low', 'high'])
df['farm'] = 'a'
df['fruit'] = arange(0,4)
results = df.groupby(['farm', 'fruit']).apply(giveMeSomeRows)
df
low high farm fruit
0 0 1 a 0
1 2 3 a 1
2 4 5 a 2
3 6 7 a 3
results
farm fruit
a 0 [0.176124290969, 0.459726835079, 0.999564934689]
1 [2.42920143009, 2.37484506501, 2.41474002256]
2 [4.78918572452, 4.25916442343, 4.77440617104]
3 [6.53831891152, 6.23242754976, 6.75141668088]
If instead you want a dataframe, you can update the function to
def giveMeSomeRows(group):
return pandas.DataFrame(random.uniform(low=group.low, high=group.high, size=N))
results
0
farm fruit
a 0 0 0.281088
1 0.020348
2 0.986269
1 0 2.642676
1 2.194996
2 2.650600
2 0 4.545718
1 4.486054
2 4.027336
3 0 6.550892
1 6.363941
2 6.702316
I have a pandas DataFrame like following.
df = pandas.DataFrame(np.random.randn(5,5),columns=['1','2','3','4','5'])
1 2 3 4 5
0 0.877455 -1.215212 -0.453038 -1.825135 0.440646
1 1.640132 -0.031353 1.159319 -0.615796 0.763137
2 0.132355 -0.762932 -0.909496 -1.012265 -0.695623
3 -0.257547 -0.844019 0.143689 -2.079521 0.796985
4 2.536062 -0.730392 1.830385 0.694539 -0.654924
I need to get row and column indexes for following three groups. (In my original dataset there are no negative values)
value is greater than 2.0
value is between 1.0 - 2.0
value is less than 1.0
For e.g for "value is greater than 2.0" it should return [1,4]. I have tried using this which gives a boolean result.
df.values > 2
You can use np.where on the boolean result to extract the indices:
import numpy as np
import pandas as pd
df = pd.DataFrame(np.random.randn(5,5),columns=['1','2','3','4','5'])
condition = df.values > 2
print np.column_stack(np.where(condition))
For a df like this,
1 2 3 4 5
0 0.057347 0.722251 0.263292 -0.168865 -0.111831
1 -0.765375 1.040659 0.272883 -0.834273 -0.126997
2 -0.023589 0.046002 1.206445 0.381532 -1.219399
3 2.290187 2.362249 -0.748805 -1.217048 -0.973749
4 0.100084 0.671120 -0.211070 0.903264 -0.312815
Output:
[[3 0]
[3 1]]
Or get a list of row-column index pairs if necessary:
print map(list, np.column_stack(np.where(condition)))
Output:
[[3,0], [3,1]]
I have a dataframe, something like:
foo bar qux
0 a 1 3.14
1 b 3 2.72
2 c 2 1.62
3 d 9 1.41
4 e 3 0.58
and I would like to add a 'total' row to the end of dataframe:
foo bar qux
0 a 1 3.14
1 b 3 2.72
2 c 2 1.62
3 d 9 1.41
4 e 3 0.58
5 total 18 9.47
I've tried to use the sum command but I end up with a Series, which although I can convert back to a Dataframe, doesn't maintain the data types:
tot_row = pd.DataFrame(df.sum()).T
tot_row['foo'] = 'tot'
tot_row.dtypes:
foo object
bar object
qux object
I would like to maintain the data types from the original data frame as I need to apply other operations to the total row, something like:
baz = 2*tot_row['qux'] + 3*tot_row['bar']
Update June 2022
pd.append is now deprecated. You could use pd.concat instead but it's probably easier to use df.loc['Total'] = df.sum(numeric_only=True), as Kevin Zhu commented. Or, better still, don't modify the data frame in place and keep your data separate from your summary statistics!
Append a totals row with
df.append(df.sum(numeric_only=True), ignore_index=True)
The conversion is necessary only if you have a column of strings or objects.
It's a bit of a fragile solution so I'd recommend sticking to operations on the dataframe, though. eg.
baz = 2*df['qux'].sum() + 3*df['bar'].sum()
df.loc["Total"] = df.sum()
works for me and I find it easier to remember. Am I missing something?
Probably wasn't possible in earlier versions.
I'd actually like to add the total row only temporarily though.
Adding it permanently is good for display but makes it a hassle in further calculations.
Just found
df.append(df.sum().rename('Total'))
This prints what I want in a Jupyter notebook and appears to leave the df itself untouched.
New Method
To get both row and column total:
import numpy as np
import pandas as pd
df = pd.DataFrame({'a': [10,20],'b':[100,200],'c': ['a','b']})
df.loc['Column_Total']= df.sum(numeric_only=True, axis=0)
df.loc[:,'Row_Total'] = df.sum(numeric_only=True, axis=1)
print(df)
a b c Row_Total
0 10.0 100.0 a 110.0
1 20.0 200.0 b 220.0
Column_Total 30.0 300.0 NaN 330.0
Use DataFrame.pivot_table with margins=True:
import pandas as pd
data = [('a',1,3.14),('b',3,2.72),('c',2,1.62),('d',9,1.41),('e',3,.58)]
df = pd.DataFrame(data, columns=('foo', 'bar', 'qux'))
Original df:
foo bar qux
0 a 1 3.14
1 b 3 2.72
2 c 2 1.62
3 d 9 1.41
4 e 3 0.58
Since pivot_table requires some sort of grouping (without the index argument, it'll raise a ValueError: No group keys passed!), and your original index is vacuous, we'll use the foo column:
df.pivot_table(index='foo',
margins=True,
margins_name='total', # defaults to 'All'
aggfunc=sum)
Voilà!
bar qux
foo
a 1 3.14
b 3 2.72
c 2 1.62
d 9 1.41
e 3 0.58
total 18 9.47
Alternative way (verified on Pandas 0.18.1):
import numpy as np
total = df.apply(np.sum)
total['foo'] = 'tot'
df.append(pd.DataFrame(total.values, index=total.keys()).T, ignore_index=True)
Result:
foo bar qux
0 a 1 3.14
1 b 3 2.72
2 c 2 1.62
3 d 9 1.41
4 e 3 0.58
5 tot 18 9.47
Building on JMZ answer
df.append(df.sum(numeric_only=True), ignore_index=True)
if you want to continue using your current index you can name the sum series using .rename() as follows:
df.append(df.sum().rename('Total'))
This will add a row at the bottom of the table.
This is the way that I do it, by transposing and using the assign method in combination with a lambda function. It makes it simple for me.
df.T.assign(GrandTotal = lambda x: x.sum(axis=1)).T
Building on answer from Matthias Kauer.
To add row total:
df.loc["Row_Total"] = df.sum()
To add column total,
df.loc[:,"Column_Total"] = df.sum(axis=1)
New method [September 2022]
TL;DR:
Just use
df.style.concat(df.agg(['sum']).style)
for a solution that won't change you dataframe, works even if you have an "sum" in your index, and can be styled!
Explanation
In pandas 1.5.0, a new method named .style.concat() gives you the ability to display several dataframes together. This is a good way to show the total (or any other statistics), because it is not changing the original dataframe, and works even if you have an index named "sum" in your original dataframe.
For example:
import pandas as pd
df = pd.DataFrame([[1, 2, 3], [4, 5, 6]], columns=['A', 'B', 'C'])
df.style.concat(df.agg(['sum']).style)
and it will return a formatted table that is visible in jupyter as this:
Styling
with a little longer code, you can even make the last row look different:
df.style.concat(
df.agg(['sum']).style
.set_properties(**{'background-color': 'yellow'})
)
to get:
see other ways to style (such as bold font, or table lines) in the docs
Following helped for me to add a column total and row total to a dataframe.
Assume dft1 is your original dataframe... now add a column total and row total with the following steps.
from io import StringIO
import pandas as pd
#create dataframe string
dfstr = StringIO(u"""
a;b;c
1;1;1
2;2;2
3;3;3
4;4;4
5;5;5
""")
#create dataframe dft1 from string
dft1 = pd.read_csv(dfstr, sep=";")
## add a column total to dft1
dft1['Total'] = dft1.sum(axis=1)
## add a row total to dft1 with the following steps
sum_row = dft1.sum(axis=0) #get sum_row first
dft1_sum=pd.DataFrame(data=sum_row).T #change it to a dataframe
dft1_sum=dft1_sum.reindex(columns=dft1.columns) #line up the col index to dft1
dft1_sum.index = ['row_total'] #change row index to row_total
dft1.append(dft1_sum) # append the row to dft1
Actually all proposed solutions render the original DataFrame unusable for any further analysis and can invalidate following computations, which will be easy to overlook and could lead to false results.
This is because you add a row to the data, which Pandas cannot differentiate from an additional row of data.
Example:
import pandas as pd
data = [1, 5, 6, 8, 9]
df = pd.DataFrame(data)
df
df.describe()
yields
0
0
1
1
5
2
6
3
8
4
9
0
count
5
mean
5.8
std
3.11448
min
1
25%
5
50%
6
75%
8
max
9
After
df.loc['Totals']= df.sum(numeric_only=True, axis=0)
the dataframe looks like this
0
0
1
1
5
2
6
3
8
4
9
Totals
29
This looks nice, but the new row is treated as if it was an additional data item, so df.describe will produce false results:
0
count
6
mean
9.66667
std
9.87252
min
1
25%
5.25
50%
7
75%
8.75
max
29
So: Watch out! and apply this only after doing all other analyses of the data or work on a copy of the DataFrame!
When the "totals" need to be added to an index column:
totals = pd.DataFrame(df.sum(numeric_only=True)).transpose().set_index(pd.Index({"totals"}))
df.append(totals)
e.g.
(Pdb) df
count min bytes max bytes mean bytes std bytes sum bytes
row_0 837200 67412.0 368733992.0 2.518989e+07 5.122836e+07 2.108898e+13
row_1 299000 85380.0 692782132.0 2.845055e+08 2.026823e+08 8.506713e+13
row_2 837200 67412.0 379484173.0 8.706825e+07 1.071484e+08 7.289354e+13
row_3 239200 85392.0 328063972.0 9.870446e+07 1.016989e+08 2.361011e+13
row_4 59800 67292.0 383487021.0 1.841879e+08 1.567605e+08 1.101444e+13
row_5 717600 112309.0 379483824.0 9.687554e+07 1.103574e+08 6.951789e+13
row_6 119600 664144.0 358486985.0 1.611637e+08 1.171889e+08 1.927518e+13
row_7 478400 67300.0 593141462.0 2.824301e+08 1.446283e+08 1.351146e+14
row_8 358800 215002028.0 327493141.0 2.861329e+08 1.545693e+07 1.026645e+14
row_9 358800 202248016.0 321657935.0 2.684668e+08 1.865470e+07 9.632590e+13
(Pdb) totals = pd.DataFrame(df.sum(numeric_only=True)).transpose()
(Pdb) totals
count min bytes max bytes mean bytes std bytes sum bytes
0 4305600.0 418466685.0 4.132815e+09 1.774725e+09 1.025805e+09 6.365722e+14
(Pdb) totals = pd.DataFrame(df.sum(numeric_only=True)).transpose().set_index(pd.Index({"totals"}))
(Pdb) totals
count min bytes max bytes mean bytes std bytes sum bytes
totals 4305600.0 418466685.0 4.132815e+09 1.774725e+09 1.025805e+09 6.365722e+14
(Pdb) df.append(totals)
count min bytes max bytes mean bytes std bytes sum bytes
row_0 837200.0 67412.0 3.687340e+08 2.518989e+07 5.122836e+07 2.108898e+13
row_1 299000.0 85380.0 6.927821e+08 2.845055e+08 2.026823e+08 8.506713e+13
row_2 837200.0 67412.0 3.794842e+08 8.706825e+07 1.071484e+08 7.289354e+13
row_3 239200.0 85392.0 3.280640e+08 9.870446e+07 1.016989e+08 2.361011e+13
row_4 59800.0 67292.0 3.834870e+08 1.841879e+08 1.567605e+08 1.101444e+13
row_5 717600.0 112309.0 3.794838e+08 9.687554e+07 1.103574e+08 6.951789e+13
row_6 119600.0 664144.0 3.584870e+08 1.611637e+08 1.171889e+08 1.927518e+13
row_7 478400.0 67300.0 5.931415e+08 2.824301e+08 1.446283e+08 1.351146e+14
row_8 358800.0 215002028.0 3.274931e+08 2.861329e+08 1.545693e+07 1.026645e+14
row_9 358800.0 202248016.0 3.216579e+08 2.684668e+08 1.865470e+07 9.632590e+13
totals 4305600.0 418466685.0 4.132815e+09 1.774725e+09 1.025805e+09 6.365722e+14
Since i generally want to do this at the very end as to avoid breaking the integrity of the dataframe (right before printing). I created a summary_rows_cols method which returns a printable dataframe:
def summary_rows_cols(df: pd.DataFrame,
column_sum: bool = False,
column_avg: bool = False,
column_median: bool = False,
row_sum: bool = False,
row_avg: bool = False,
row_median: bool = False
) -> pd.DataFrame:
ret = df.copy()
if column_sum: ret.loc['Sum'] = df.sum(numeric_only=True, axis=0)
if column_avg: ret.loc['Avg'] = df.mean(numeric_only=True, axis=0)
if column_median: ret.loc['Median'] = df.median(numeric_only=True, axis=0)
if row_sum: ret.loc[:, 'Sum'] = df.sum(numeric_only=True, axis=1)
if row_median: ret.loc[:, 'Avg'] = df.mean(numeric_only=True, axis=1)
if row_avg: ret.loc[:, 'Median'] = df.median(numeric_only=True, axis=1)
ret.fillna('-', inplace=True)
return ret
This allows me to enter a generic (numeric) df and get a summarized output such as:
a b c Sum Median
0 1 4 7 12 4
1 2 5 8 15 5
2 3 6 9 18 6
Sum 6 15 24 - -
from:
data = {
'a': [1, 2, 3],
'b': [4, 5, 6],
'c': [7, 8, 9]
}
df = pd.DataFrame(data)
printable = summary_rows_cols(df, row_sum=True, column_sum=True, row_median=True)