Home > Mobile >  iterate over a df and multiply the values by the values of another df
iterate over a df and multiply the values by the values of another df

Time:05-14

My df1 looks like this:It contains 3 unique project id.The date starts on 01-01-22 and ends on 01-12-28

id date p50 p90
apv1 01-01-22 1000 1000
apv2 01-01-22 1000 1000
tsso 01-01-22 1202 2005
apv1 01-02-22 1000 2000
apv2 01-02-22 1400 5000
tsso 01-02-22 200 1000
. . . .
. . . .
. . . .
apv1 01-11-28 305 400
apv2 01-11-28 300 200
tsso 01-11-28 250 499
apv1 01-12-28 100 290
apv2 01-12-28 145 650
tsso 01-12-28 299 179

df2 looks like this:

mth month pct
01 january 0.042
02 febuary 0.055
03 march 0.089
04 april 0.097
05 may 0.10
06 june 0.11
07 july 0.127
08 august 0.114
09 sept 0.091
10 oct 0.068
11 nov 0.043
12 dec 0.038

I want to multiply each value p50 of df1 by the value pct of df2 with mth or month value equals to the month of the date in df1. Similarly for p90.

The Final df should look like this:

id date p50 p90
apv1 01-01-22 1000*0.042 1000*0.042
apv2 01-01-22 1000*0.042 1000*0.042
tsso 01-01-22 1202*0.042 2005*0.042
apv1 01-02-22 1000*0.055 2000*0.055
apv2 01-02-22 1400*0.055 5000*0.055
tsso 01-02-22 200*0.055 1000*0.055
. . . .
. . . .
. . . .
apv1 01-11-28 305*0.043 400*0.043
apv2 01-11-28 300*0.043 200*0.043
tsso 01-11-28 250*0.043 499*0.043
apv1 01-12-28 100*0.038 290*0.038
apv2 01-12-28 145*0.038 650*0.038
tsso 01-12-28 299*0.038 179*0.038

CodePudding user response:

Simply assign month column using the date column. Then merge both data frames by month. Finally, run multiplication with needed columns:

final_df = (
    df1.assign(mth = lambda x: pd.to_datetime(x["date"]).dt.strftime("%m"))
       .merge(df2, on="mth")
       .assign(
           p50 = lambda x: x["p50"].mul(x["pct"]),
           p90 = lambda x: x["p90"].mul(x["pct"])
       )
)

CodePudding user response:

You should be able to use map to modify you data in place:

s = df2.set_index('mth')['pct']
pct = pd.to_datetime(df1['date']).dt.strftime('%m').map(s)

df1[['p50', 'p90']] *= pct
  • Related