Is there a way in Pandas to subtract two values that are in the same column that have the same name?

Question

Here is a snippet of a dataframe I'm trying to analyze. What I want to do is simply subtract FP_FLOW FORMATTED_ENTRY values from D8_FLOW FORMATTED_ENTRY values only if the X_LOT_NAME is the same. For example, in the X_LOT_NAME column you can see MPACZX2. The D8_FLOW FORMATTED_ENTRY is 12.3%. The FP_FLOW FORMATTED_ENTRY value is 7.8%. The difference between the two would be 4.5%. I want to apply this logic across the whole data set

Answer 1

Is this what you are looking for?

df.groupby(['x_lot'])['value'].diff()

0     NaN
1     NaN
2    -5.0
3     8.0
4    -3.0
5     NaN
6   -10.0
Name: value, dtype: float64

This is the data i used to get the above results

    x_lot   type    value
0   mpaczw1 fp  21
1   mpaczw2 d8  12
2   mpaczw2 fp  7
3   mpaczw2 d8  15
4   mpaczw2 fp  12
5   mpaczw3 d8  21
6   mpaczw3 fp  11

Answer 2

it is advisable to first convert your data into a format where the values to be added / subtracted are in the same row, and after that subtract / add the corresponding oclumns. You can do this using pd.pivot-table . The below example will demonstrate this using a sample dataframe similar to what you've shared:

wanted_data

    X_LOT_NAME  SPEC_TYPE   FORMATTED_ENTRY
0   a   FP_FLOW     1
1   a   D8_FLOW     2
2   c   FP_FLOW     3
3   c   D8_FLOW     4

pivot_data  = pd.pivot_table(wanted_data,values='FORMATTED_ENTRY',index='X_LOT_NAME',columns='SPEC_TYPE')
pivot_data
SPEC_TYPE   D8_FLOW     FP_FLOW
X_LOT_NAME      
a   2   1
c   4   3

After this step, the resultant pivot_data contains the same data, but the columns are D8_FLOW and FP_FLOW , with X_LOT_NAME as the index. Now you can get the intended value in a new column using:

pivot_data['DIFF'] =  pivot_data['D8_FLOW'] - pivot_data['FP_FLOW']