Selecting Columns to fill based on numbers of items in row - pandas

Question

So I have 4 columns in a dataframe: W, X, Y, Z.
I have a CSV file that has some rows with some having 4 items, 3 items, and 2 items per row.
I am using:
frame = pd.read_csv("file_example.csv", names = [ 'W', 'X', 'Y', 'Z'])
Is there a way to make it so that the rows with 3 fill in W X and Z skipping over Y and leaving then NAN? And similarly make it so that the rows with 2 items fill in W and Z, skipping over X and Y? As it is now it just fills in the first columns it comes across.
In other words, is there a way to pick and choose which columns a row will fill up based on the number of items in the row?
Thanks.
Edit:
Input (corresponding to the output):
2,seafood,21418
2,stews,24126
2,seafood,23287
2,sandwiches,17429
and
4,6237
4,30815
4,5321
4,49248
Trying the method below, I put 100 test lines each of 4,3,2 item rows.
Sample part of output:
3 item line:
2 seafood 21418.0 21418
2 stews 24126.0 24126
2 seafood 23287.0 23287
2 sandwiches 17429.0 17429
2 item line:
4 6237 NaN 6237
4 30815 NaN 30815
4 5321 NaN 5321
4 49248 NaN 49248
The z is filling correctly, but the NaNs are not masking over.
Edit 2: Did not assign the new dataframe to a variable. Solution works.

Answer 1

import numpy as np
import pandas as pd
df = pd.read_csv('test.csv', names=['W', 'X', 'Y', 'Z'])
df
Out:
W X Y Z
0 10 Blue 20160809.0 203.0
1 12 Red 20160810.0 4578.0
2 9 Red 3094.0 NaN
3 15 Yellow 109.0 NaN
4 1 86 NaN NaN
5 5 9384 NaN NaN
6 56 3490 NaN NaN
Record the positions of NaNs:
nans = df.isnull().values
Fill Z column:
df['Z'] = df['Z'].fillna(df['Y'].fillna(df['X']))
Shift NaNs to the left:
df.mask(np.roll(nans, -1), np.nan)
Out:
W X Y Z
0 10 Blue 20160809.0 203
1 12 Red 20160810.0 4578
2 9 Red NaN 3094
3 15 Yellow NaN 109
4 1 NaN NaN 86
5 5 NaN NaN 9384
6 56 NaN NaN 3490