How to assign unique key with reference with another dataframe values in pandas?

user_17068495 0
Viewed 129

I want to create column which take reference from df1 and map to df2. df1:

    Id         Value
    [1,3,4]    Furn
    [1,3,4]    Furn
    [1,7,9]    SYS
    [1,2,3]    YSU
    [1,3,4]    Furn
    [1,7,9]    SYS

df2:

 Id        Val2
[1,3,5,4]   Kcv
[1,3,9,4]   Kcm
[1,0,9,7]   jks
[5,7,9,1]   kja
[1,2,9,7]   kla
[3,2,0,8]   ola
[1,2,4,5]   opa

df_output:

    Id        Val2    df1_reference
    [1,3,5,4]   Kcv   [1,3,4]       -- Every element present in df1.Id
    [1,3,9,4]   Kcm   [1,3,4]       -- Every element present in df1.Id
    [1,0,9,7]   jks   [1,7,9]       -- Every element present in df1.Id
    [5,7,9,1]   kja   [1,7,9]       -- Every element present in df1.Id
    [1,2,9,7]   kla   [1,7,9]       -- Every element present in df1.Id
    [3,2,0,8]   ola   NaN
    [1,2,4,5]   opa   NaN

I need to create "df1_reference" columns based on the element present in df.Id, if df.Id list element present in df2.Id then map with df.Id.

pandas python
1 Answers

You need to use a loop here. You can take advantage of the set/frozenset methods:

ids = set(df1['Id'].apply(frozenset)) # use tuple instead of frozenset
                                      # if order matters

df2['df1_reference'] = [next((x for x in ids if x.issubset(s)), np.nan)
                        for s in df2['Id']]

Or if you want to keep the original objects of df1['Id']:

ids = {frozenset(x): x for x in df1['Id']} # use tuple instead if order matters

df2['df1_reference'] = [next((v for k,v in ids.items() if k.issubset(s)), np.nan)
                        for s in df2['Id']]

output:

             Id Val2 df1_reference
0  [1, 3, 5, 4]  Kcv     [1, 3, 4]
1  [1, 3, 9, 4]  Kcm     [1, 3, 4]
2  [1, 0, 9, 7]  jks     [1, 7, 9]
3  [5, 7, 9, 1]  kja     [1, 7, 9]
4  [1, 2, 9, 7]  kla     [1, 7, 9]
5  [3, 2, 0, 8]  ola           NaN
6  [1, 2, 4, 5]  opa           NaN
user_163434640
Related