CODEWITHSUNDEEP
pythonpandasdictionary
jmaval
jmaval

Reputation: 313

Merged list of dictionaries

I have the below list containing dictionaries, which themselves contain dataframes, that I would like to merge. 

ls = [{'varID': 101, 'varDesc': 'Description of variable 101', 'values': df101},
      {'varID': 102, 'varDesc': 'Description of variable 102', 'values': df102},
      {'varID': 103, 'varDesc': 'Description of variable 103', 'values': df103}]

Where df101, df102, and df103 are all x by 1 dataframes in shape with date values as the index (named 'refPer'), and the column name of 'values' in all the dataframes. Note: that x is not the same size for all the dateframes.

I would like to merge the dataframes and have attempted the following: 

mergedDF = reduce(lambda x,y: pd.merge(x.get('values'),y.get('values'), how = 'outer', on = 'refPer'), ls)

However, I get the following error: 

ValueError: can not merge DataFrame with instance of type <class 'NoneType'>

Any help would be appreciated.

Thank you in advance

Upvotes: 0

Views: 51

Answers (1)

busybear
busybear

Reputation: 10580

You can use concat to join Dataframes of unequal length. You can also provide the parameter keys to give a name to each dataframe you are concatenating. You'll create a multiindex, but you can drop the original level if you don't want it:

df = pd.concat([x['values'] for x in ls], axis=1, keys=[x['varID'] for x in ls])
df.columns = df.columns.droplevel(1)  # removes 'original' level

Upvotes: 1

Related Questions