I have a DataFrame as shown below

DateTime                 Order     Val1         Val2        Val3    Category    Item    Quantity    Cost
0   01-01-2000 00:00    Offline 906437.46   507604.062  217544.598  Appliances  Oven    654     1108.79
1   01-01-2000 00:00    Offline 71512.00    40046.720   17162.880   Clothing    Sweater 2554    22.40
2   01-01-2000 00:00    Offline 74987.50    41993.000   17997.000   Clothing    Shirt   4285    14.00
3   01-01-2000 00:00    Offline 1097057.80  614354.020  263294.580  Gadgets     Ipad     590    1487.54
4   01-01-2000 00:00    Offline 725866.48   406489.104  174209.616  Gadgets   HeadPhone 1384    419.58
5   01-01-2000 00:00    Offline 177827.28   99586.242   42679.818   Gadgets     Ipod    1059    134.34
6   01-01-2000 00:00    Offline 88381.21    49491.211   21210.519   Utensils FryingPan 1619     43.67
7   01-01-2000 00:00    Offline 40555.44    22708.728   9748.872    Utensils    Spoon   1656    19.59
8   01-01-2000 00:00    Online  150752.00   84421.120   36180.480   Clothing    Jacket  2692    44.80
9   01-01-2000 00:00    Online  58861.59    32961.313   14126.277   Utensils    Pot     841     55.99
10  02-01-2000 00:00    Offline 752395.25   421326.425  180574.575  Appliances  Oven    475    1267.19
11  02-01-2000 00:00    Offline 105920.00   59315.200   25420.800   Clothing    Jacket  1655    51.20
12  02-01-2000 00:00    Offline 79540.00    44542.400   19089.600   Clothing    Shirt   3977    16.00
13  02-01-2000 00:00    Offline 511879.06   286651.078  122850.462  Gadgets   HeadPhone 854     479.51
14  02-01-2000 00:00    Offline 62826.73    35181.559   15077.811   Utensils  FryingPan 1007    49.91
15  02-01-2000 00:00    Offline 39801.78    22287.006   9551.574    Utensils    Spoon   1422    22.39
16  02-01-2000 00:00    Online  77792.00    43563.520   18670.080   Clothing    Sweater 2431    25.60
17  02-01-2000 00:00    Online  554638.05   310597.308  133113.132  Gadgets     Ipad    261    1700.04
18  02-01-2000 00:00    Online  173870.46   97368.726   41729.454   Gadgets     Ipod    906    153.53
19  02-01-2000 00:00    Online  60632.42    33953.094   14551.326   Utensils    Pot     758     63.99

I want to write a lag function in which I can lag the numerical columns by x number of days, which I plan to use on different datasets. The function will get the category columns, date column and the numerical columns I want to lag in the data as input. I want the output to look like:

DateTime                 Order     Val1         Val2        Val3    Category    Item    Quantity    Cost  Quantity_lag
0   01-01-2000 00:00    Offline 906437.46   507604.062  217544.598  Appliances  Oven    654     1108.79     Nan
1   01-01-2000 00:00    Offline 71512.00    40046.720   17162.880   Clothing    Sweater 2554    22.40       Nan
2   01-01-2000 00:00    Offline 74987.50    41993.000   17997.000   Clothing    Shirt   4285    14.00       Nan
3   01-01-2000 00:00    Offline 1097057.80  614354.020  263294.580  Gadgets     Ipad     590    1487.54     Nan
4   01-01-2000 00:00    Offline 725866.48   406489.104  174209.616  Gadgets   HeadPhone 1384    419.58      Nan
5   01-01-2000 00:00    Offline 177827.28   99586.242   42679.818   Gadgets     Ipod    1059    134.34      Nan
6   01-01-2000 00:00    Offline 88381.21    49491.211   21210.519   Utensils FryingPan 1619     43.67       Nan
7   01-01-2000 00:00    Offline 40555.44    22708.728   9748.872    Utensils    Spoon   1656    19.59       Nan
8   01-01-2000 00:00    Online  150752.00   84421.120   36180.480   Clothing    Jacket  2692    44.80       Nan
9   01-01-2000 00:00    Online  58861.59    32961.313   14126.277   Utensils    Pot     841     55.99       Nan
10  02-01-2000 00:00    Offline 752395.25   421326.425  180574.575  Appliances  Oven    475    1267.19      Nan
11  02-01-2000 00:00    Offline 105920.00   59315.200   25420.800   Clothing    Jacket  1655    51.20       Nan
12  02-01-2000 00:00    Offline 79540.00    44542.400   19089.600   Clothing    Shirt   3977    16.00       Nan
13  02-01-2000 00:00    Offline 511879.06   286651.078  122850.462  Gadgets   HeadPhone 854     479.51      Nan
14  02-01-2000 00:00    Offline 62826.73    35181.559   15077.811   Utensils  FryingPan 1007    49.91       Nan
15  02-01-2000 00:00    Offline 39801.78    22287.006   9551.574    Utensils    Spoon   1422    22.39       Nan
16  02-01-2000 00:00    Online  77792.00    43563.520   18670.080   Clothing    Sweater 2431    25.60       Nan
17  02-01-2000 00:00    Online  554638.05   310597.308  133113.132  Gadgets     Ipad    261    1700.04      Nan
18  02-01-2000 00:00    Online  173870.46   97368.726   41729.454   Gadgets     Ipod    906    153.53       Nan
19  02-01-2000 00:00    Online  60632.42    33953.094   14551.326   Utensils    Pot     758     63.99       Nan

I tried using the following code:

groupby_columns = [datecolumn] + category_columns
df[datecolumn] = pd.to_datetime(df[datecolumn], format = '%d-%m-%Y %H:%M')
grp_col = df.merge(df.assign(DateTime = df[datecolumn]+pd.Timedelta(days=lag)), on=groupby_columns, how='left', suffixes = ['','_lag'])

which gives me the desired output but it lags all the numerical columns that are present inside the data. Is there a better method for doing this?