suppose I have two dataframes:
import pandas
....
....
test1 = pandas.DataFrame([1,2,3,4,5])
....
....
test2 = pandas.DataFrame([4,2,1,3,7])
....
I tried test1.append(test2) but it is the equivalent of R's rbind.
How can I combine the two as two columns of a dataframe similar to the cbind function in R?
test3 = pd.concat([test1, test2], axis=1)
test3.columns = ['a','b']
(But see the detailed answer by #feng-mai, below)
There is a key difference between concat(axis = 1) in pandas and cbind() in R:
concat attempts to merge/align by index. There is no concept of index in a R dataframe. If the two pandas dataframes' indexes are misaligned, the results are different from cbind (even if they have the same number of rows). You need to either make sure the indexes align or drop/reset the indexes.
Example:
import pandas as pd
test1 = pd.DataFrame([1,2,3,4,5])
test1.index = ['a','b','c','d','e']
test2 = pd.DataFrame([4,2,1,3,7])
test2.index = ['d','e','f','g','h']
pd.concat([test1, test2], axis=1)
0 0
a 1.0 NaN
b 2.0 NaN
c 3.0 NaN
d 4.0 4.0
e 5.0 2.0
f NaN 1.0
g NaN 3.0
h NaN 7.0
pd.concat([test1.reset_index(drop=True), test2.reset_index(drop=True)], axis=1)
0 1
0 1 4
1 2 2
2 3 1
3 4 3
4 5 7
pd.concat([test1.reset_index(), test2.reset_index(drop=True)], axis=1)
index 0 0
0 a 1 4
1 b 2 2
2 c 3 1
3 d 4 3
4 e 5 7
Related
I have a margin table
item margin
0 a 3
1 b 4
2 c 5
and an item table
item sequence
0 a 1
1 a 2
2 a 3
3 b 1
4 b 2
5 c 1
6 c 2
7 c 3
I want to join the two table so that the margin will only be joined to the product with maximum sequence number, the desired outcome is
item sequence margin
0 a 1 NaN
1 a 2 NaN
2 a 3 3.0
3 b 1 NaN
4 b 2 4.0
5 c 1 NaN
6 c 2 NaN
7 c 3 5.0
How to achieve this?
Below is the code for margin and item table
import pandas as pd
df_margin=pd.DataFrame({"item":["a","b","c"],"margin":[3,4,5]})
df_item=pd.DataFrame({"item":["a","a","a","b","b","c","c","c"],"sequence":[1,2,3,1,2,1,2,3]})
One option would be to merge then replace extra values with NaN via Series.where:
new_df = df_item.merge(df_margin)
new_df['margin'] = new_df['margin'].where(
new_df.groupby('item')['sequence'].transform('max').eq(new_df['sequence'])
)
Or with loc:
new_df = df_item.merge(df_margin)
new_df.loc[new_df.groupby('item')['sequence']
.transform('max').ne(new_df['sequence']), 'margin'] = np.NAN
Another option would be to assign a temp column to both frames df_item with True where the value is maximal, and df_margin is True everywhere then merge outer and drop the temp column:
new_df = (
df_item.assign(
t=df_item
.groupby('item')['sequence']
.transform('max')
.eq(df_item['sequence'])
).merge(df_margin.assign(t=True), how='outer').drop('t', 1)
)
Both produce new_df:
item sequence margin
0 a 1 NaN
1 a 2 NaN
2 a 3 3.0
3 b 1 NaN
4 b 2 4.0
5 c 1 NaN
6 c 2 NaN
7 c 3 5.0
You could do:
df_item.merge(df_item.groupby('item')['sequence'].max().\
reset_index().merge(df_margin), 'left')
item sequence margin
0 a 1 NaN
1 a 2 NaN
2 a 3 3.0
3 b 1 NaN
4 b 2 4.0
5 c 1 NaN
6 c 2 NaN
7 c 3 5.0
Breakdown:
df_new = df_item.groupby('item')['sequence'].max().reset_index().merge(df_margin)
df_item.merge(df_new, 'left')
I have a long form dataframe that contains multiple samples and time points for each subject. The number of samples and timepoint can vary, and the days between time points can also vary:
test_df = pd.DataFrame({"subject_id":[1,1,1,2,2,3],
"sample":["A", "B", "C", "D", "E", "F"],
"timepoint":[19,11,8,6,2,12],
"time_order":[3,2,1,2,1,1]
})
subject_id sample timepoint time_order
0 1 A 19 3
1 1 B 11 2
2 1 C 8 1
3 2 D 6 2
4 2 E 2 1
5 3 F 12 1
I need to figure out a way to generalize grouping this dataframe by subject_id and putting all samples and time points on the same row, in time order.
DESIRED OUTPUT:
subject_id sample1 timepoint1 sample2 timepoint2 sample3 timepoint3
0 1 C 8 B 11 A 19
1 2 E 2 D 6 null null
5 3 F 12 null null null null
Pivot gets me close, but I'm stuck on how to proceed from there:
test_df = test_df.pivot(index=['subject_id', 'sample'],
columns='time_order', values='timepoint')
Use DataFrame.set_index with DataFrame.unstack for pivoting, sorting MultiIndex in columns, flatten it and last convert subject_id to column:
df = (test_df.set_index(['subject_id', 'time_order'])
.unstack()
.sort_index(level=[1,0], axis=1))
df.columns = df.columns.map(lambda x: f'{x[0]}{x[1]}')
df = df.reset_index()
print (df)
subject_id sample1 timepoint1 sample2 timepoint2 sample3 timepoint3
0 1 C 8.0 B 11.0 A 19.0
1 2 E 2.0 D 6.0 NaN NaN
2 3 F 12.0 NaN NaN NaN NaN
a=test_df.iloc[:,:3].groupby('subject_id').last().add_suffix('1')
b=test_df.iloc[:,:3].groupby('subject_id').nth(-2).add_suffix('2')
c=test_df.iloc[:,:3].groupby('subject_id').nth(-3).add_suffix('3')
pd.concat([a, b,c], axis=1)
sample1 timepoint1 sample2 timepoint2 sample3 timepoint3
subject_id
1 C 8 B 11.0 A 19.0
2 E 2 D 6.0 NaN NaN
3 F 12 NaN NaN NaN NaN
I have a dataset with lots of variables. So I've extracted the numeric ones:
numeric_columns = transposed_df.select_dtypes(np.number)
Then I want to replace all 0 values for 0.0001
transposed_df[numeric_columns.columns] = numeric_columns.where(numeric_columns.eq(0, axis=0), 0.0001)
And here is the first problem. This line is not replacing the 0 values with 0.0001, but is replacing all non zero values with 0.0001.
Also after this (replacing the 0 values by 0.0001) I want to replace all values there are less than the 1th quartile of the row to -1 and leave the others as they were. But I am not managing how.
To answer your first question
In [36]: from pprint import pprint
In [37]: pprint( numeric_columns.where.__doc__)
('\n'
'Replace values where the condition is False.\n'
'\n'
'Parameters\n'
'----------\n'
because of that your all the values except 0 are getting replaced
Use DataFrame.mask and for second condition compare by DataFrame.quantile:
transposed_df = pd.DataFrame({
'A':list('abcdef'),
'B':[0,0.5,4,5,5,4],
'C':[7,8,9,4,2,3],
'D':[1,3,0,7,1,0],
'E':[5,3,6,9,2,4],
'F':list('aaabbb')
})
numeric_columns = transposed_df.select_dtypes(np.number)
m1 = numeric_columns.eq(0)
m2 = numeric_columns.lt(numeric_columns.quantile(q=0.25, axis=1), axis=0)
transposed_df[numeric_columns.columns] = numeric_columns.mask(m1, 0.0001).mask(m2, -1)
print (transposed_df)
A B C D E F
0 a -1.0 7 1.0 5 a
1 b -1.0 8 3.0 3 a
2 c 4.0 9 -1.0 6 a
3 d 5.0 -1 7.0 9 b
4 e 5.0 2 -1.0 2 b
5 f 4.0 3 -1.0 4 b
EDIT:
from scipy.stats import zscore
print (transposed_df[numeric_columns.columns].apply(zscore))
B C D E
0 -2.236068 0.570352 -0.408248 0.073521
1 0.447214 0.950586 0.408248 -0.808736
2 0.447214 1.330821 -0.816497 0.514650
3 0.447214 -0.570352 2.041241 1.838037
4 0.447214 -1.330821 -0.408248 -1.249865
5 0.447214 -0.950586 -0.816497 -0.367607
EDIT1:
transposed_df = pd.DataFrame({
'A':list('abcdef'),
'B':[0,1,1,1,1,1],
'C':[1,8,9,4,2,3],
'D':[1,3,0,7,1,0],
'E':[1,3,6,9,2,4],
'F':list('aaabbb')
})
numeric_columns = transposed_df.select_dtypes(np.number)
from scipy.stats import zscore
df1 = pd.DataFrame(numeric_columns.apply(zscore, axis=1).tolist(),index=transposed_df.index)
transposed_df[numeric_columns.columns] = df1
print (transposed_df)
A B C D E F
0 a -1.732051 0.577350 0.577350 0.577350 a
1 b -1.063410 1.643452 -0.290021 -0.290021 a
2 c -0.816497 1.360828 -1.088662 0.544331 a
3 d -1.402136 -0.412393 0.577350 1.237179 b
4 e -1.000000 1.000000 -1.000000 1.000000 b
5 f -0.632456 0.632456 -1.264911 1.264911 b
I want to do groupby, shift and cumsum which seems pretty trivial task but still banging my head over the result I'm getting. Can someone please tell what am I doing wrong. All the results I found online shows the same or the same variation of what I am doing. Below is my implementation.
temp = pd.DataFrame(data=[['a',1],['a',1],['a',1],['b',1],['b',1],['b',1],['c',1],['c',1]], columns=['ID','X'])
temp['transformed'] = temp.groupby('ID')['X'].cumsum().shift()
print(temp)
ID X transformed
0 a 1 NaN
1 a 1 1.0
2 a 1 2.0
3 b 1 3.0
4 b 1 1.0
5 b 1 2.0
6 c 1 3.0
7 c 1 1.0
This is wrong because the actual or what I am looking for is as below:
ID X transformed
0 a 1 NaN
1 a 1 1.0
2 a 1 2.0
3 b 1 NaN
4 b 1 1.0
5 b 1 2.0
6 c 1 NaN
7 c 1 1.0
Thanks a lot in advance.
You could use transform() to feed the separate groups that are created at each level of groupby into the cumsum() and shift() methods.
temp['transformed'] = \
temp.groupby('ID')['X'].transform(lambda x: x.cumsum().shift())
ID X transformed
0 a 1 NaN
1 a 1 1.0
2 a 1 2.0
3 b 1 NaN
4 b 1 1.0
5 b 1 2.0
6 c 1 NaN
7 c 1 1.0
For more info on transform() please see here:
https://jakevdp.github.io/PythonDataScienceHandbook/03.08-aggregation-and-grouping.html#Transformation
https://pandas.pydata.org/pandas-docs/version/0.22/groupby.html#transformation
You need using apply , since one function is under groupby object which is cumsum another function shift is for all df
temp['transformed'] = temp.groupby('ID')['X'].apply(lambda x : x.cumsum().shift())
temp
Out[287]:
ID X transformed
0 a 1 NaN
1 a 1 1.0
2 a 1 2.0
3 b 1 NaN
4 b 1 1.0
5 b 1 2.0
6 c 1 NaN
7 c 1 1.0
While working on this problem, as the DataFrame size grows, using lambdas on transform starts to get very slow. I found out that using some DataFrameGroupBy methods (like cumsum and shift instead of lambdas are much faster.
So here's my proposed solution, creating a 'temp' column to save the cumsum for each ID and then shifting in a different groupby:
df['temp'] = df.groupby("ID")['X'].cumsum()
df['transformed'] = df.groupby("ID")['temp'].shift()
df = df.drop(columns=["temp"])
I am new to Python and Pandas , can someone help me with below report.
I want to report difference of N columns and create new columns with difference value, is it possible to make it dynamic as I have more than 30 columns. (Columns are fixed numbers, rows values can change)
A and B can be Alpha numeric
Use join with sub for difference of DataFrames:
#if columns are strings, first cast it
df1 = df1.astype(int)
df2 = df2.astype(int)
#if first columns are not indices
#df1 = df1.set_index('ID')
#df2 = df2.set_index('ID')
df = df1.join(df2.sub(df1).add_prefix('sum'))
print (df)
A B sumA sumB
ID
0 10 2.0 5 3.0
1 11 3.0 6 5.0
2 12 4.0 7 5.0
Or similar:
df = df1.join(df2.sub(df1), rsuffix='sum')
print (df)
A B Asum Bsum
ID
0 10 2.0 5 3.0
1 11 3.0 6 5.0
2 12 4.0 7 5.0
Detail:
print (df2.sub(df1))
A B
ID
0 5 3.0
1 6 5.0
2 7 5.0
IIUC
df1[['C','D']]=(df2-df1)[['A','B']]
df1
Out[868]:
ID A B C D
0 0 10 2.0 5 3.0
1 1 11 3.0 6 5.0
2 2 12 4.0 7 5.0
df1.assign(B=0)
Out[869]:
ID A B C D
0 0 10 0 5 3.0
1 1 11 0 6 5.0
2 2 12 0 7 5.0
The 'ID' column should really be an index. See the Pandas tutorial on indexing for why this is a good idea.
df1 = df1.set_index('ID')
df2 = df2.set_index('ID')
df = df1.copy()
df[['C', 'D']] = df2 - df1
df['B'] = 0
print(df)
outputs
A B C D
ID
0 10 0 5 3.0
1 11 0 6 5.0
2 12 0 7 5.0