pythonpandasdataframepython-3.11

How to avoid fragmentation warning when moving a column to position 0?

single statement of

df.insert(0, 'time', df.pop('time'))

on large dataframe causing

warning: PerformanceWarning: DataFrame is highly fragmented.

is there away to move a column to position ZERO to avoid this warning.

Solution

  • Consider this dataframe:

       column1  column2  time
0        1        2     3
1        4        5     6

    You can reorder the columns and put the time column first like this:

    df = df[["time", *df.columns.difference(["time"])]]
print(df)

    Prints:

       time  column1  column2
0     3        1        2
1     6        4        5