New pandas data frame at given index values

Asked Apr 09 '19 at 10:30

Active Apr 09 '19 at 10:36

Viewed 42 times

I have the following pandas data frame:

import pandas as pd
timestamps = np.arange(1, 53)
quantities = np.arange(1, 53)
data = dict(quantities=quantities)
df = pd.DataFrame(data=data, columns=data.keys(), index=timestamps)
df.rename_axis('timestamps', inplace=True)

I'm trying to get change the data frame to different values:

timestamps_new = [1, 1, 1, 2, 2, 2]
df = df.loc[timestamps_new]

This works, but I get the following warning:

FutureWarning: Passing list-likes to .loc or [] with any missing label will raise
KeyError in the future, you can use .reindex() as an alternative.

Why does this happen? I also get a KeyError when I use:

df = df.loc[timestamps_new, 'index']

df = df.loc[timestamps_new, 'timestamps']

Thanks for any help understanding this.

asked Apr 09 '19 at 10:30

ajrlewis

2,968
3
33
67

It's not an error, just a `FutureWarning`. You can suppress them: `import warnings; warnings.simplefilter(action='ignore', category=FutureWarning)` – Chris Adams Apr 09 '19 at 10:33
also, `loc` is failing because you do not have a column with name `index` or `timestamps`. Try `df.loc[timestamps_new, 'quantities']` – Chris Adams Apr 09 '19 at 10:34
@ChrisA What if you want all columns, rather than just `quantities`? – ajrlewis Apr 09 '19 at 10:40
then just `df.loc[timestamps_new]` or `df.loc[timestamps_new, :]` will do – Chris Adams Apr 09 '19 at 10:41
@ChrisA Thanks. Both options still raise the warning though. – ajrlewis Apr 09 '19 at 10:45
see https://stackoverflow.com/questions/15777951/how-to-suppress-pandas-future-warning – Chris Adams Apr 09 '19 at 10:47
@ChrisA Thanks, though the end of the warning says it'll raise an error in future, so would rather solve, than suppress, the issue. – ajrlewis Apr 09 '19 at 10:48
1

Just use the suggestion? `df.reindex(timestamps_new)` – Chris Adams Apr 09 '19 at 10:50
1

@ChrisA Yep ... :-/ Thanks – ajrlewis Apr 09 '19 at 10:53
which version of pandas you are using?. i am using 0.21.0 and i am not getting any warning df = df.loc[timestamps_new] – Chirag Apr 09 '19 at 12:41

New pandas data frame at given index values

0 Answers0