如何动态创建和合并数据框？

Question

I have 10 csv files which are huge in size. I would like to

1) read those files

2) create dataframes (with filename as dataframe name)

3) left outer join all of them based on joining keys which are given. POIU and BVCX have only one common column to merge which is A . Please note file 'ABCDE' is the base df.. All other dataframe should be left outer joined with this 'ABCDE' df. It is possible that there are other common keys as well. But I would like to join based on keys A and B whichever of these two exist

I was able to do the first two steps as shown below

filenames = sorted(glob.glob('*.csv'))
df_list=[]
for f in filenames:
    print(f)
    t = vars()['df'+ f = pd1.read_csv(f,low_memory=False)
    df_list.append(t)

But I am stuck on how to left outer join all this on the fly and create one final dataframe which will be named as df_final ?

Answer 1

If the joining keys are all the shared keys that exists in the files, you don't have to do anything special to change from one joined key to 2. You can merge them in the loop with:

for f in filenames:

    # YOUR CODE WITH WHATEVER YOU DO WITH IT
    print(f)
    t = vars()['df'+ f = pd1.read_csv(f,low_memory=False)
    df_list.append(t)

    # THE CODE FOR MERGING THE DFS
    cur_df = pd1.read_csv(f, low_memory=False)
    try:
        JKeys = ['A', 'B'] if 'B' in cur_df.columns else ['A']:
        df_final = df_final.merge(cur_df, right_on=JKeys, how='left')
    except:
        df_final = cur_df