Rahul shrivastava Rahul shrivastava - 3 months ago 20
Python Question

How should I subtract two dataframes and in Pandas and diplay the required output?

My table looks like this:

In [82]:df.head()
Out[82]:
MatDoc MatYr MvT Material Plnt SLoc Batch Customer AmountLC Amount ... PO MatYr.1 MatDoc.1 Order ProfitCtr SLED/BBD PstngDate EntryDate Time Username
0 4912693062 2015 551 100062 HDC2 0001 5G30MC1A11 NaN 9.03 9.06 ... NaN NaN NaN NaN IN1165B085 26.01.2016 01.08.2015 01.08.2015 01:13:16 O33462
1 4912693063 2015 501 166 HDC2 0004 NaN NaN 0.00 0.00 ... NaN NaN NaN NaN IN1165B085 NaN 01.08.2015 01.08.2015 01:13:17 O33462
2 4912693320 2015 551 101343 HDC2 0001 5G28MC1A11 NaN 53.73 53.72 ... NaN NaN NaN NaN IN1165B085 25.01.2016 01.08.2015 01.08.2015 01:16:30 O33462


Here, I need to group by data on
Order
column and sum only
AmountLC
column.Then I need to check for the
Order
column values such that it should be present in both
MvT101group
and
MvT102group
. and if an
Order
matches in both sets of data then I need to subtract
MvT102group
from
MvT101group
. and display

Order|Plnt|Material|Batch|Sum101=SumofMvt101ofAmountLC|Sum102=SumofMvt102ofAmountLC|(Sum101-Sum102)/100


What I have done is first I made new df containing only 101 and 102:
Mvt101
and
MvT102


MvT101 = df.loc[df['MvT'] == 101]


MvT102 = df.loc[df['MvT'] == 102]


Then I grouped it by
Order
and got the sum value for the column

MvT101group = MvT101.groupby('Order', sort=True)


In [76]:
MvT101group[['AmountLC']].sum()
Out[76]:
Order AmountLC
1127828 16348566.88
1127829 22237710.38
1127830 29803745.65
1127831 30621381.06
1127832 33926352.51


MvT102group = MvT102.groupby('Order', sort=True)


In [77]:
MvT102group[['AmountLC']].sum()
Out[77]:
Order AmountLC
1127830 53221.70
1127831 651475.13
1127834 67442.16
1127835 2477494.17
1128622 218743.14


After this I am not able to understand how should I write my query.
Please ask me any further details if you want.Here is the CSV file from where I am working Link

Answer

Hope I understood the question correctly. After grouping both groups as you did:

MvT101group = MvT101.groupby('Order',sort=True).sum()
MvT102group = MvT102.groupby('Order',sort=True).sum()

You can update the columns' names for both groups:

MvT101group.columns = MvT101group.columns.map(lambda x: str(x) + '_101')
MvT102group.columns = MvT102group.columns.map(lambda x: str(x) + '_102')

Then merge all 3 tables so that you will have all 3 columns in the main table:

df = df.merge(MvT101group, left_on=['Order'], right_index=True, how='left')
df = df.merge(MvT102group, left_on=['Order'], right_index=True, how='left')

And then you can add the calculated column:

df['calc'] = (df['Order_101']-df['Order_102']) / 100