Replace all non NaN values with the initial value per row-CodePudding

I have the following dataframe:

Periods            0      1      2      3      4      5      6      7     8     9     10    11    12   13   14   15   16   17   18   19
Item 1           89.0   89.0   88.0   86.0   73.0   73.0   67.0    NaN   NaN   NaN   NaN   NaN   NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
Item 2          197.0  197.0  157.0  157.0  156.0  156.0  153.0  153.0   3.0   3.0   2.0   NaN   NaN  NaN  NaN  NaN  NaN  NaN  NaN  NaN
Item 3          105.0  105.0   98.0   47.0   40.0   21.0   20.0   15.0  13.0  11.0  11.0  11.0  10.0  7.0  7.0  5.0  5.0  3.0  2.0  1.0

It has several NaNs, and columns from a range between 0 - 20. What i wanted, is to replicate the first column to the next columns, until this value is not NaN. Let me show you an example of what it should look like:

Periods                0      1      2      3      4      5      6      7      8      9     10     11     12     13     14     15     16     17     18     19
Item 1               89.0   89.0   89.0   89.0   89.0   89.0   89.0    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN
Item 2               197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN
Item 3               105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0

But I can't figure out a way to do this. I wish i could post some code, but all I tried was to search for those methods: backfill, ffill. None of them looked like what I needed.

Any ideas? Thanks in advance.

CodePudding user response：

I am assuming here that "Periods" is the index

You can use mask on the columns with df.notna() as mask, and the first column (df.iloc[:, 0]) as replacement values:

df.mask(df.notna(), df.iloc[:, 0], axis=0)

output:

             0      1      2      3      4      5      6      7      8      9     10     11     12     13     14     15     16     17     18     19
Periods                                                                                                                                            
Item 1    89.0   89.0   89.0   89.0   89.0   89.0   89.0    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN
Item 2   197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0  197.0    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN    NaN
Item 3   105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0  105.0

CodePudding user response：

Here's another idea, just for the record: using map:

df.iloc[0] = df.iloc[0].notna().map({True: df.iloc[0, 0]})

CodePudding user response：

I will do mul

df.notna().mul(df.iloc[:,0],axis=0).mask(df.isna())