How to partition a dataframe into dataframes with no missing and contiguous indices?

Suppose I have the following DataFrame

index  Value
  0     'A'
  1     'B'
  2    pd.NA
  3    pd.NA
  4     'C'
  5    pd.NA

Then, I’m looking for a function that returns the next two DataFrames:

First:

index  Value
  0     'A'
  1     'B'

Second:

index  Value
  4     'C'

>Solution :

You can build a mask of NAs (with isna), form groups of successive non-NAs with cumsum and split with groupby:

m = df['Value'].isna()

dfs = [g for _,g in df[~m].groupby(m.cumsum())]

Output:

[   index Value
 0      0     A
 1      1     B,
    index Value
 4      4     C]

Intermediates:

   index Value   isna      ~  cumsum
0      0     A  False   True       0
1      1     B  False   True       0
2      2   NaN   True  False       1
3      3   NaN   True  False       2
4      4     C  False   True       2
5      5   NaN   True  False       3