CODEWITHSUNDEEP
pythonsqlpandas
Sebastian Zeki
Sebastian Zeki

Reputation: 6874

Pandas: Merge two dataframes on id and matching date range within +/- certain days

The aim:

I want to merge two dataframes on the basis of the unique number and the date matching within +/-7 days

The data:

df1

Number         Report         DateDone
1       some words      13/1/2021
1               more stuff      21/8/2021
44      balbla          11/4/2020
2       gobbledy bla    01/03/2019
44      rara rasputin   13/10/2021
44      tree frogs      11/10/2010

df2

Number         Report             DateDone
1       hocum poklum       11/1/2021
1       mjimmeny cricket   21/8/2021
44      it wasnt me        11/2/2020
2       its not really     6/03/2019
44      im innocent        12/10/2021
44      bullfrogs          11/01/2010

The intended result

Number.df1     Report.df1   DateDone.df1     Number.df2    Report.df2     DateDone.df2
1              some words    13/1/2021              1          hocum poklum      11/1/2021
1              more stuff    21/8/2021              1          jimmeny cricket   21/8/2021
2              gobbledy bla  01/03/2019             2          its not really    6/03/2019
44             rara rasputin 13/10/2021             44         im innocent       12/10/2021

I was going to use a sql merge similar to one I found here but I am having difficulty knowing how to merge on the number and a date range. Do I need to calculate the 7 days before and after the DateDone in df1? Surely there is a more efficient way than having to calculate two new columns first?

qry = '''
    select  
        df1.DateDone_start TermStart,
        df1.DateDone_end TermEnd,
        df2.DateDone df2Start,
        df1.Number,
        df2.Number
    from
        df1 join df2 on
        date between df1.DateDone_start and df1.DateDone_end join df1 on
        df1.Number = df2.Number
    '''
df = pd.read_sql_query(qry, conn)

Upvotes: 1

Views: 387

Answers (2)

SeaBean
SeaBean

Reputation: 23217

You can use .merge() on Number and then filter using .loc the condition where DateDone.df2 is .between() DateDone.df1 +/- 7 days by using +/-pd.DateOffset(days=7), as follows:

df1['DateDone'] = pd.to_datetime(df1['DateDone'], dayfirst=True)
df2['DateDone'] = pd.to_datetime(df2['DateDone'], dayfirst=True)

df_merge = df1.merge(df2, on='Number', suffixes=('.df1', '.df2'))

result = df_merge.loc[
             df_merge['DateDone.df2'].between(
                 df_merge['DateDone.df1'] - pd.DateOffset(days=7), 
                 df_merge['DateDone.df1'] + pd.DateOffset(days=7))]

Result:

print(result)



    Number     Report.df1 DateDone.df1        Report.df2 DateDone.df2
0        1     some words   2021-01-13      hocum poklum   2021-01-11
3        1     more stuff   2021-08-21  mjimmeny cricket   2021-08-21
8       44  rara rasputin   2021-10-13       im innocent   2021-10-12
13       2   gobbledy bla   2019-03-01    its not really   2019-03-06

Upvotes: 3

Henry Ecker
Henry Ecker

Reputation: 35636

Try merge then filter out rows that are within 7 days:

new_df = df1.merge(df2, on='Number', suffixes=('.df1', '.df2'))
new_df = new_df[
    abs(new_df['DateDone.df1'] - new_df['DateDone.df2']) <= pd.Timedelta(days=7)
    ]

new_df:

    Number     Report.df1 DateDone.df1        Report.df2 DateDone.df2
0        1     some words   2021-01-13      hocum poklum   2021-01-11
3        1     more stuff   2021-08-21  mjimmeny cricket   2021-08-21
8       44  rara rasputin   2021-10-13       im innocent   2021-10-12
13       2   gobbledy bla   2019-03-01   its not really    2019-03-06

Convert to 'DateDone' to DateTime for both Frames if not already done:

df1['DateDone'] = pd.to_datetime(df1['DateDone'], format='%d/%m/%Y')
df2['DateDone'] = pd.to_datetime(df2['DateDone'], format='%d/%m/%Y')

Get the duration between two datetimes

new_df['DateDone.df1'] - new_df['DateDone.df2']

0        2 days
1     -220 days
2      222 days
3        0 days
4       60 days
5     -549 days
6     3743 days
7      610 days
8        1 days
9     4293 days
10   -3410 days
11   -4019 days
12     273 days
13      -5 days
dtype: timedelta64[ns]

Apply abs to remove directionality from durations and compare to the desired duration:

abs(new_df['DateDone.df1'] - new_df['DateDone.df2']) <= pd.Timedelta(days=7)

Use this index to determine which rows to keep:

0      True
1     False
2     False
3      True
4     False
5     False
6     False
7     False
8      True
9     False
10    False
11    False
12    False
13     True
dtype: bool

Upvotes: 0

Related Questions