Copy Previous Row Values when Current Value is 0 (Zero)

Question

My Dataframe as below:

13664567    74.3    SELL    1581566.4   2378211.6   12:07:59
13665406    74.3    0       0           0           0
13665406    0       0       0           0           0
13665406    74.3    0       0           0           0
13667531    74.3    0       0           0           0
13667531    74.3    0       0           0           0
13672281    74.3    0       0           0           0
13672591    74.3    0       0           0           0
13672591    74.3    BUY     2388742.8   1590276     12:08:01
13681398    74.3    0       0           0           0
13681398    74.3    0       0           0           0
13684573    74.3    0       0           0           0
13685574    74.3    0       0           0           0
13685574    74.3    0       0           0           0
13688544    0       0       0           0           0
13689596    74.3    0       0           0           0
13689596    74.3    0       0           0           0
13705735    74.3    0       0           0           0
13706035    74.3    0       0           0           0
13712130    74.3    SELL    1583219.2   2408620.8   12:08:01
13712130    74.3    0       0           0           0
13715699    74.3    0       0           0           0
13720809    74.3    0       0           0           0
13726310    74.3    0       0           0           0
13726310    74.3    0       0           0           0
13726410    74.3    0       0           0           0

I want to fill the 0 (Zero) values of last 4 Columns with previous row values.

I want dataframe like as below:

13664567    74.3    SELL    1581566.4    2378212    12:07:59
13665406    74.3    SELL    1581566.4    2378212    12:07:59
13665406    0       SELL    1581566.4    2378212    12:07:59
13665406    74.3    SELL    1581566.4    2378212    12:07:59
13667531    74.3    SELL    1581566.4    2378212    12:07:59
13667531    74.3    SELL    1581566.4    2378212    12:07:59
13672281    74.3    SELL    1581566.4    2378212    12:07:59
13672591    74.3    SELL    1581566.4    2378212    12:07:59
13672591    74.3    BUY     2388742.8    1590276    12:08:01
13681398    74.3    BUY     2388742.8    1590276    12:08:01
13681398    74.3    BUY     2388742.8    1590276    12:08:01
13684573    74.3    BUY     2388742.8    1590276    12:08:01
13685574    74.3    BUY     2388742.8    1590276    12:08:01
13685574    74.3    BUY     2388742.8    1590276    12:08:01
13688544    0       BUY     2388742.8    1590276    12:08:01
13689596    74.3    BUY     2388742.8    1590276    12:08:01
13689596    74.3    BUY     2388742.8    1590276    12:08:01
13705735    74.3    BUY     2388742.8    1590276    12:08:01
13706035    74.3    BUY     2388742.8    1590276    12:08:01
13712130    74.3    SELL    1583219.2    2408621    12:08:01
13712130    74.3    SELL    1583219.2    2408621    12:08:01
13715699    74.3    SELL    1583219.2    2408621    12:08:01
13720809    74.3    SELL    1583219.2    2408621    12:08:01
13726310    74.3    SELL    1583219.2    2408621    12:08:01
13726310    74.3    SELL    1583219.2    2408621    12:08:01
13726410    74.3    SELL    1583219.2    2408621    12:08:01

I want to make the changes only to last four columns and copy the previous row values to next row until the New Value is available.

Please help.

Answer 1

Use update with select last 4 columns by iloc, replace by missing values and forward fill last previous values:

df.update(df.iloc[:, -4:].replace([0,'0'], np.nan).ffill())

Or use mask for replace 0:

df.update(df.iloc[:, -4:].mask(lambda x: x.isin([0, '0'])).ffill())
#if no string 0
#df.update(df.iloc[:, -4:].mask(lambda x: x == 0)).ffill())

print (df)
           a     b     c          d          e         f
0   13664567  74.3  SELL  1581566.4  2378211.6  12:07:59
1   13665406  74.3  SELL  1581566.4  2378211.6  12:07:59
2   13665406  74.3  SELL  1581566.4  2378211.6  12:07:59
3   13665406  74.3  SELL  1581566.4  2378211.6  12:07:59
4   13667531  74.3  SELL  1581566.4  2378211.6  12:07:59
5   13667531  74.3  SELL  1581566.4  2378211.6  12:07:59
6   13672281  74.3  SELL  1581566.4  2378211.6  12:07:59
7   13672591  74.3  SELL  1581566.4  2378211.6  12:07:59
8   13672591  74.3   BUY  2388742.8  1590276.0  12:08:01
9   13681398  74.3   BUY  2388742.8  1590276.0  12:08:01
10  13681398  74.3   BUY  2388742.8  1590276.0  12:08:01
11  13684573  74.3   BUY  2388742.8  1590276.0  12:08:01
12  13685574  74.3   BUY  2388742.8  1590276.0  12:08:01
13  13685574  74.3   BUY  2388742.8  1590276.0  12:08:01
14  13688544  74.3   BUY  2388742.8  1590276.0  12:08:01
15  13689596  74.3   BUY  2388742.8  1590276.0  12:08:01
16  13689596  74.3   BUY  2388742.8  1590276.0  12:08:01
17  13705735  74.3   BUY  2388742.8  1590276.0  12:08:01
18  13706035  74.3   BUY  2388742.8  1590276.0  12:08:01
19  13712130  74.3  SELL  1583219.2  2408620.8  12:08:01
20  13712130  74.3  SELL  1583219.2  2408620.8  12:08:01
21  13715699  74.3  SELL  1583219.2  2408620.8  12:08:01
22  13720809  74.3  SELL  1583219.2  2408620.8  12:08:01
23  13726310  74.3  SELL  1583219.2  2408620.8  12:08:01
24  13726310  74.3  SELL  1583219.2  2408620.8  12:08:01
25  13726410  74.3  SELL  1583219.2  2408620.8  12:08:01

Answer 2

You can use mask + ffill:

values = df.iloc[:, -4:]
df.iloc[:, -4:] = values.mask(values.isin([0, '0'])).ffill()