I have the following DataFrame:

I would like to write a function that will return the number of NaN values before the first non-NaN value. Given the above example, the function should return the value 2.

I tried to solve my problem using this question, but it did not help me much.

Edit: The values always start with a NaN. If the column is all NaN, the function should return the column length.

CodePudding user response：

You could use isna to get True/1 on the NaN values and cumprod to get rid of all values that follow a non-NaN. Then sum:

df['y'].isna().cumprod().sum()

output: 2

CodePudding user response：

You can use first_valid_index.

df.y.first_valid_index()
> 2

This grabs the index of the first non-NaN value. By default we don't need to sum if the index starts from NaN.

CodePudding user response：

Use Series.isna with Series.cummin and count Trues by sum:

s = df['y'].isna().cummin().sum()
print (s)
2