Develop Reference

How to add dataframe to multiindex dataframe at specific location

I'm organizing data from separate files into one portable, multiindex
dataframe, with multiindex ("A", "B", "C"). Some of the info is gathered
from the filenames read in, and should populate the "A", and "B", of the
multiindex. "C" should take the form of the index of the file read in.
The columns should take the form of the columns read in.
Let's say the files read in become:
df1
0 1 2 3 4
0 0 9 9 8 5
1 0 8 2 1 2
2 9 1 6 4 3
3 1 4 1 4 4
4 5 4 6 6 2
df2
0 1 2 3 4
0 4 5 0 7 3
1 8 2 9 1 0
2 5 9 1 6 6
3 4 1 4 6 5
4 3 0 0 8 8
How do I get to this end result:
multiindex_df
0 1 2 3 4
A B C
1 1 0 0 9 9 8 5
1 0 8 2 1 2
2 9 1 6 4 3
3 1 4 1 4 4
4 5 4 6 6 2
1 2 0 4 5 0 7 3
1 8 2 9 1 0
2 5 9 1 6 6
3 4 1 4 6 5
4 3 0 0 8 8
Starting from:
import pandas as pd
import numpy as np
multiindex_df = pd.DataFrame(
index=pd.MultiIndex.from_arrays(
[[], [], []], names=["A", "B", "C"]))
df1 = pd.DataFrame(np.random.randint(10, size=(5, 5)))
df1_a = 1
df1_b = 1
df2 = pd.DataFrame(np.random.randint(10, size=(5, 5)))
df2_a = 1
df2_b = 2
breakpoint()
This is what I have in mind, but gives a key error:
multiindex_df.loc[(df1_a, df1_b, slice(None))] = df1
multiindex_df.loc[(df2_a, df2_b, slice(None))] = df2

You could do this as follows:
multiindex_df = pd.concat([df1, df2], keys=[1,2])
multiindex_df = pd.concat([multiindex_df], keys=[1])
multiindex_df.index.names = ['A','B','C']
print(multiindex_df)
0 1 2 3 4
A B C
1 1 0 0 9 9 8 5
1 0 8 2 1 2
2 9 1 6 4 3
3 1 4 1 4 4
4 5 4 6 6 2
2 0 4 5 0 7 3
1 8 2 9 1 0
2 5 9 1 6 6
3 4 1 4 6 5
4 3 0 0 8 8
Alternatively, you could do it like below:
# collect your dfs inside a dict
dfs = {'1': df1, '2': df2}
# create list for index tuples
multi_index = []
for idx, val in enumerate(dfs):
for x in dfs[val].index:
# append tuple to row, e.g (1,1,0), (1,1,1) etc.
multi_index.append((1,idx+1,x))
# concat your dfs
multiindex_df_two = pd.concat([df1, df2])
# create multiindex from tuples, and add names
multiindex_df_two.index = pd.MultiIndex.from_tuples(multi_index, names=['A','B','C'])
# check
multiindex_df.equals(multiindex_df_two) # True

Ouroboros' answer made
me realize that instead of trying to fit the read-in file dfs into a
formatted df, the cleaner solution is to format each individual file df,
then concat.
In order to do that, I have to re-format the file df indices into a
multiindex, which prompted this question and
answer.
Having that down, and using Ouroboros'
answer, the solution
becomes:
import pandas as pd
import numpy as np
df1 = pd.DataFrame(np.random.randint(10, size=(5, 5)))
df1_a = 1
df1_b = 1
df1.index = pd.MultiIndex.from_product(
[[df1_a], [df1_b], df1.index], names=["A", "B", "C"])
df2 = pd.DataFrame(np.random.randint(10, size=(5, 5)))
df2_a = 1
df2_b = 2
df2.index = pd.MultiIndex.from_product(
[[df2_a], [df2_b], df2.index], names=["A", "B", "C"])
multiindex_df = pd.concat([df1, df2])
Which is obviously well suited for a loop.
Output:
df1
0 1 2 3 4
0 5 3 1 1 3
1 8 9 7 5 6
2 8 6 6 7 7
3 3 4 9 7 2
4 3 2 1 6 2
df2
0 1 2 3 4
0 5 0 6 9 3
1 7 5 5 9 6
2 2 1 9 6 3
3 9 4 3 7 0
4 5 9 5 9 6
multiiindex_df
0 1 2 3 4
A B C
1 1 0 5 3 1 1 3
1 8 9 7 5 6
2 8 6 6 7 7
3 3 4 9 7 2
4 3 2 1 6 2
2 0 5 0 6 9 3
1 7 5 5 9 6
2 2 1 9 6 3
3 9 4 3 7 0
4 5 9 5 9 6

Sort a subset of columns of a pandas dataframe alphabetically by column name

I'm having trouble finding the solution to a fairly simple problem.
I would like to alphabetically arrange certain columns of a pandas dataframe that has over 100 columns (i.e. so many that I don't want to list them manually).
Example df:
import pandas as pd
subject = [1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,4,4,4,4,4,4]
timepoint = [1,2,3,4,5,6,1,2,3,4,5,6,1,2,4,1,2,3,4,5,6]
c = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
d = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
a = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
b = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
df = pd.DataFrame({'subject':subject,
'timepoint':timepoint,
'c':c,
'd':d,
'a':a,
'b':b})
df.head()
subject timepoint c d a b
0 1 1 2 2 2 2
1 1 2 3 3 3 3
2 1 3 4 4 4 4
3 1 4 5 5 5 5
4 1 5 6 6 6 6
How could I rearrange the column names to generate a df.head() that looks like this:
subject timepoint a b c d
0 1 1 2 2 2 2
1 1 2 3 3 3 3
2 1 3 4 4 4 4
3 1 4 5 5 5 5
4 1 5 6 6 6 6
i.e. keep the first two columns where they are and then alphabetically arrange the remaining column names.
Thanks in advance.

You can split your your dataframe based on column names, using normal indexing operator [], sort alphabetically the other columns using sort_index(axis=1), and concat back together:
>>> pd.concat([df[['subject','timepoint']],
df[df.columns.difference(['subject', 'timepoint'])]\
.sort_index(axis=1)],ignore_index=False,axis=1)
subject timepoint a b c d
0 1 1 2 2 2 2
1 1 2 3 3 3 3
2 1 3 4 4 4 4
3 1 4 5 5 5 5
4 1 5 6 6 6 6
5 1 6 7 7 7 7
6 2 1 3 3 3 3
7 2 2 4 4 4 4
8 2 3 1 1 1 1
9 2 4 2 2 2 2
10 2 5 3 3 3 3
11 2 6 4 4 4 4
12 3 1 5 5 5 5
13 3 2 4 4 4 4
14 3 4 5 5 5 5
15 4 1 8 8 8 8
16 4 2 4 4 4 4
17 4 3 5 5 5 5
18 4 4 6 6 6 6
19 4 5 2 2 2 2
20 4 6 3 3 3 3

Specify the first two columns you want to keep (or determine them from the data), then sort all of the other columns. Use .loc with the correct list to then "sort" the DataFrame.
import numpy as np
first_cols = ['subject', 'timepoint']
#first_cols = df.columns[0:2].tolist() # OR determine first two
other_cols = np.sort(df.columns.difference(first_cols)).tolist()
df = df.loc[:, first_cols+other_cols]
print(df.head())
subject timepoint a b c d
0 1 1 2 2 2 2
1 1 2 3 3 3 3
2 1 3 4 4 4 4
3 1 4 5 5 5 5
4 1 5 6 6 6 6

You can try getting the dataframe columns as a list, rearrange them, and assign it back to the dataframe using df = df[cols]
import pandas as pd
subject = [1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,4,4,4,4,4,4]
timepoint = [1,2,3,4,5,6,1,2,3,4,5,6,1,2,4,1,2,3,4,5,6]
c = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
d = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
a = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
b = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
df = pd.DataFrame({'subject':subject,
'timepoint':timepoint,
'c':c,
'd':d,
'a':a,
'b':b})
cols = df.columns.tolist()
cols = cols[:2] + sorted(cols[2:])
df = df[cols]

Combining columns within the same df Python/Pandas

I'm new to the programming world and can't figure out how to concatenate columns in pandas. I'm not looking to join these columns, but rather stack them on top of each other.
This is the code I have so far:
import pandas as pd
import numpy as np
df = pd.read_excel("C:\\Users\\Kit Wesselhoeft\\Documents\\NEM\\Northend Manufacturing_deletecol.xlsx")
print(df)
df = pd.concat(['A','A'])
print(df)
image here
I want to combine all the columns so that all the A's sit on top of each other, Same with the B's - E's.
How can I do this? Am I missing something?

It looks like you are looking for "append":
import numpy as np
import pandas as pd
df = pd.DataFrame(np.random.randint(1,10, (3,2)),columns=list('AB'))
df2 = pd.DataFrame(np.random.randint(1,10, (3,2)),columns=list('AB'))
df3=df.append(df2)
In [2]: df3
Out[2]:
A B
0 7 6
1 8 3
2 2 1
0 2 2
1 1 3
2 5 5

If you are certain column ordering is consistent and tiled [A,B,C,A,B,C...], then you can create a new DataFrame by reshaping the old. Otherwise safer alternatives exist with pd.wide_to_long which uses the actual column names.
Sample Data
import numpy as np
import pandas as pd
np.random.seed(123)
df = pd.DataFrame(np.random.randint(1, 10, (3, 15)),
columns=list('BACDE')*3)
# B A C D E B A C D E B A C D E
#0 3 3 7 2 4 7 2 1 2 1 1 4 5 1 1
#1 5 2 8 4 3 5 8 3 5 9 1 8 4 5 7
#2 2 6 7 3 2 9 4 6 1 3 7 3 5 5 7
Reshape
cols = pd.unique(df.columns) # Preserves Order
pd.DataFrame(df.values.reshape(-1, len(cols)), columns=cols)
# B A C D E
#0 3 3 7 2 4
#1 7 2 1 2 1
#2 1 4 5 1 1
#3 5 2 8 4 3
#4 5 8 3 5 9
#5 1 8 4 5 7
#6 2 6 7 3 2
#7 9 4 6 1 3
#8 7 3 5 5 7
pd.wide_to_long
Useful when your columns are not in the same tiling order, or if you have more of some than others. Requires you to modify the column names by adding _N for which occurrence of the column it is.
cols = pd.unique(df.columns)
s = pd.Series(df.columns).groupby(df.columns).cumcount()
df.columns = [f'{col}_{N}' for col,N in zip(df.columns, s)]
pd.wide_to_long(df.reset_index(), stubnames=cols, i='index', j='num', sep='_').reset_index(drop=True)
# B A C D E
#0 3 3 7 2 4
#1 5 2 8 4 3
#2 2 6 7 3 2
#3 7 2 1 2 1
#4 5 8 3 5 9
#5 9 4 6 1 3
#6 1 4 5 1 1
#7 1 8 4 5 7
#8 7 3 5 5 7

The following example is relevant when you exactly know where your columns are. Building on ALollz's code:
import numpy as np
import pandas as pd
np.random.seed(123)
df = pd.DataFrame(np.random.randint(1, 10, (3, 15)),
columns=list('BACDE')*3)
# B A C D E B A C D E B A C D E
#0 3 3 7 2 4 7 2 1 2 1 1 4 5 1 1
#1 5 2 8 4 3 5 8 3 5 9 1 8 4 5 7
#2 2 6 7 3 2 9 4 6 1 3 7 3 5 5 7
# Using iloc
df1 = df.iloc[:, :5]
df2 = df.iloc[:,5:10]
df3 = df.iloc[:,10:]
df_final= pd.concat([df1,df2,df3]).reset_index(drop=True)
Result df_final:
B A C D E
0 3 3 7 2 4
1 5 2 8 4 3
2 2 6 7 3 2
3 7 2 1 2 1
4 5 8 3 5 9
5 9 4 6 1 3
6 1 4 5 1 1
7 1 8 4 5 7
8 7 3 5 5 7

Holding a first value in a column while another column equals a value?

I would like to hold the first value in a column while another column does not equal zero. For Column B, values alternate between -1, 0, 1. For Column C, values equal any integer. The objective is holding the first value of Column C while Column B equals zero. The current DataFrame is as follows:
A B C
1 8 1 9
2 2 1 1
3 3 0 7
4 9 0 8
5 5 0 9
6 6 0 1
7 1 1 9
8 6 1 10
9 3 0 4
10 8 0 8
11 5 0 9
12 6 0 10
The resulting DataFrame should be as follows:
A B C
1 8 1 9
2 2 1 1
3 3 0 7
4 9 0 7
5 5 0 7
6 6 0 7
7 1 1 9
8 6 1 10
9 3 0 4
10 8 0 4
11 5 0 4
12 6 0 4
13 3 1 9

You need first create NaNs by condition in column C and then add values by ffill:
mask = (df['B'].shift().fillna(False)).astype(bool) | (df['B'])
df['C'] = df.loc[mask, 'C']
df['C'] = df['C'].ffill().astype(int)
print (df)
A B C
1 8 1 9
2 2 1 1
3 3 0 7
4 9 0 7
5 5 0 7
6 6 0 7
7 1 1 9
8 6 1 10
9 3 0 4
10 8 0 4
11 5 0 4
12 6 0 4
13 3 1 9
Or use where and if type of all values is integer, add astype:
mask = (df['B'].shift().fillna(False)).astype(bool) | (df['B'])
df['C'] = df['C'].where(mask).ffill().astype(int)
print (df)
A B C
1 8 1 9
2 2 1 1
3 3 0 7
4 9 0 7
5 5 0 7
6 6 0 7
7 1 1 9
8 6 1 10
9 3 0 4
10 8 0 4
11 5 0 4
12 6 0 4
13 3 1 9

pandas add variables according to variable value

Suppose the following pandas dataframe
Wafer_Id v1 v2
0 0 9 6
1 0 7 8
2 0 1 5
3 1 6 6
4 1 0 8
5 1 5 0
6 2 8 8
7 2 2 6
8 2 3 5
9 3 5 1
10 3 5 6
11 3 9 8
I want to group it according to WaferId and I would like to get something like
w
Out[60]:
Wafer_Id v1_1 v1_2 v1_3 v2_1 v2_2 v2_3
0 0 9 7 1 6 ... ...
1 1 6 0 5 6
2 2 8 2 3 8
3 3 5 5 9 1
I think that I can obtain the result with the pivot function but I am not sure of how to do it

Possible solution
oes = pd.DataFrame()
oes['Wafer_Id'] = [0,0,0,1,1,1,2,2,2,3,3,3]
oes['v1'] = np.random.randint(0, 10, 12)
oes['v2'] = np.random.randint(0, 10, 12)
oes['id'] = [0, 1, 2] * 4
oes.pivot(index='Wafer_Id', columns='id')
oes
Out[74]:
Wafer_Id v1 v2 id
0 0 8 7 0
1 0 3 3 1
2 0 8 0 2
3 1 2 5 0
4 1 4 1 1
5 1 8 8 2
6 2 8 6 0
7 2 4 7 1
8 2 4 3 2
9 3 4 6 0
10 3 9 2 1
11 3 7 1 2
oes.pivot(index='Wafer_Id', columns='id')
Out[75]:
v1 v2
id 0 1 2 0 1 2
Wafer_Id
0 8 3 8 7 3 0
1 2 4 8 5 1 8
2 8 4 4 6 7 3
3 4 9 7 6 2 1