4

Not sure why I still get the warning after I updated to .loc method suggested by the message? Is it a false alarm?

eG.loc[:,'wt']=eG.groupby(['date','BB'])['m'].transform(weightFunction)

A value is trying to be set on a copy of a slice from a DataFrame

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  if __name__ == '__main__':
Lisa
  • 4,126
  • 12
  • 42
  • 71

2 Answers2

15

I guess your eG DF is a copy of another DF...

Here is a small demo:

In [69]: df = pd.DataFrame(np.random.randint(0, 5, (10, 3)),  columns=list('abc'))

In [70]: cp = df[df.a > 0]

In [71]: cp.loc[:, 'c'] = cp.groupby('a').b.transform('sum')
c:\envs\py35\lib\site-packages\pandas\core\indexing.py:549: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  self.obj[item_labels[indexer[info_axis]]] = value

Workaround:

In [72]: cp = df[df.a > 0].copy()

In [73]: cp.loc[:, 'c'] = cp.groupby('a').b.transform('sum')

Or if you don't need original DF you can save memory:

In [74]: df = df[df.a > 0]

In [75]: df.loc[:, 'c'] = df.groupby('a').b.transform('sum')
MaxU - stand with Ukraine
  • 205,989
  • 36
  • 386
  • 419
  • I wonder why do we need to add .copy() if we are trying to get a transformed version of a dataframe? Thanks! – CathyQian Jun 14 '18 at 23:27
0

This is to avoid what is called chained indexing. Pandas Chained Index

As Marx suggested, by using the deep copy, you easily can skip this warning.

cp = df[df.a > 0].copy()

or if it is a heavy data set and you do not need the original one just replace the slice with the original.

df = df[df.a > 0]
Behrooz Hosseini
  • 61
  • 3