Pandas 单列内多个条件

Question

In the below df, I need to replace the COST A & COST B for E to 0 and replace Comment as Un reported cost , when below conditions are met -

1. E and F have the same cost for 'COST A'
2. E and F have the same cost for 'COST B'

as you can see 20 and 0.5 for E is replace with 0, as E and F have the same cost

df = pd.DataFrame([['1/1/2021','SKU_1','A','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','B','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','C','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','D','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','E','0.05','20','Calculated'],
                   ['1/1/2021','SKU_1','F','0.05','20','Actual']],
                   columns = ['MTH-YR','SKU','TYPE','COST A','COST B','COMMENT'])

Expected result,

     MTH-YR     SKU   TYPE COST A  COST B  COMMENT
0   1/1/2021    SKU_1   A   0   0   Un reported cost
1   1/1/2021    SKU_1   B   0   0   Un reported cost
2   1/1/2021    SKU_1   C   0   0   Un reported cost
3   1/1/2021    SKU_1   D   0   0   Un reported cost
4   1/1/2021    SKU_1   E   0   0   Un reported cost
5   1/1/2021    SKU_1   F   0.5 20  Actual

Answer 1

for i in range (len(df["COST A"])):
    # First check for the index of E:
    if df["TYPE"][i] == "E":
        # Then we have two conditions that will be true if E == F
        if df["COST A"][i] == df["COST A"][i+1] and df["COST B"][i] == df["COST B"][i+1]:
            # Now we replace the old values           
            df["COST A"][i] = 0
            df["COST B"][i] = 0
            df["COMMENT"][i] = "Un reported cost"

Answer 2

If this is just for this small dataset, use .iloc method.

## Change COST B for E to 0
df.iloc[4,4] = 0

## Change COMMENT for E to 'Un reported cost'
df.iloc[4,5] = 'Un reported cost'

UPDATE: I was playing around with this before going to sleep, and came across this solution on a larger data.

Let's define your data. I have added duplicate E's here.

## Create dataframe
df = pd.DataFrame([['1/1/2021','SKU_1','A','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','B','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','C','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','D','0','0','Un reported cost'],
                   ['1/1/2021','SKU_1','E','0.05','20','Calculated'],
                   ['1/1/2021','SKU_1','E','0.05','20','Calculated'],
                   ['1/1/2021','SKU_1','E','0.05','20','Calculated'],
                   ['1/1/2021','SKU_1','F','0.05','20','Actual']],
                   columns = ['MTH-YR','SKU','TYPE','COST A','COST B','COMMENT'])

Using list comprehension, we can assign values to COST B based on the conditions that you need.

df['COST B'] = ['0' if df['TYPE'][x] == 'E' else df['COST B'][x] for x in range(len(df['TYPE']))]

In essence, we're just looping through the values in TYPE to see which ones are type 'E'. When found, we're changing that value to 0. Otherwise, we're keeping the value stored in COST B . Same logic is applied for COMMENT .

df['COMMENT'] = ['Un reported cost' if df['TYPE'][x] == 'E' else df['COMMENT'][x] for x in range(len(df['TYPE']))]