+- Python Forum (https://python-forum.io)
+-- Forum: Python Coding (https://python-forum.io/forum-7.html)
+--- Forum: Data Science (https://python-forum.io/forum-44.html)
+--- Thread: pandas : insert new column (/thread-2462.html)
pandas : insert new column - vvv - Mar-18-2017
Hi All,
I am trying to create an empty dataframe 'a' and insert slices of raw data from 'df' in to columns in 'a'. I can insert a one new columns and other columns get NULL value
l=[0,1,2,3,4,5,6,7,8,9,10]
df = pd.DataFrame(l)
df.columns = ['c']
a=pd.DataFrame()
a.insert(0,0,df.c.loc[0:2])
a.insert(1,1,df.c.loc[3:6]) ## inserts only NaN
Out[134]:
0 1
0 0 NaN
1 1 NaN
2 2 NaN
Thanks
vvv
RE: pandas : insert new column - zivoni - Mar-18-2017
It is not entirely clear what you want to do, as you are trying to insert a column with length 4 into a dataframe with length 3. Pandas respects indices when you try to insert second column and tries to insert values for index values 0, 1, 2 - there are none so NaN is used.
Output:In [10]: import pandas as pd
In [11]: df = pd.DataFrame({'c': range(11)})
In [12]: a = pd.DataFrame()
In [13]: a.insert(0, 0, df.c.loc[0:2])Output:In [14]: pd.concat([a, df.c.loc[3:6]], axis=1)
Out[14]:
0 c
0 0.0 NaN
1 1.0 NaN
2 2.0 NaN
3 NaN 3.0
4 NaN 4.0
5 NaN 5.0
6 NaN 6.0Output:In [15]: pd.concat([a, df.c.loc[3:6].reset_index(drop=True)], axis=1)
Out[15]:
0 c
0 0.0 3
1 1.0 4
2 2.0 5
3 NaN 6RE: pandas : insert new column - vvv - Mar-19-2017
Thanks Zivoni. I used the concat and fix my index .
It worked.