Searching for match in pandas dataframe-CodePudding

With the following csv file

id,name,cy,in
0,MD,4,16
2,MD,10,20
3,YD,5,14
4,ZD,10,14

I have written the following code to create a new dataframe.

df = pd.read_csv('test.csv', usecols=['id', 'name', 'cy', 'in'])
df2 = pd.DataFrame(columns=['N', 'I', 'C'])
ids=[0,2,4]
for i in ids:
    row = df.loc[df['id'] == i]
    cyc = row.at[row.index[0],'cy']
    ins = row.at[row.index[0],'in']
    name = row.at[row.index[0],'name']
    if df2['N'].str.contains(name):
        print("Matched")
    else:
    new_row = {'N':name, 'I':ins, 'C':cyc}
    df_temp = pd.DataFrame([new_row])
    df2 = pd.concat([df2, df_temp], axis=0, ignore_index=True)
print(df2)

As you can see, for specified id, I first get the row from the original dataframe, df. If I couldn't find the name in the second dataframe, df2, then create a new row and add that to the second dataframe. However, on a match, I would like to add the values to the existing row. So, in the end, I expect to see:

    N   I   C
0  MD  36  14
2  ZD  14  10

That if statement has the following error though:

ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

Howe can I fix that?

CodePudding user response：

Use Series.isin for get all id in boolean indexing and then aggregate sum in names aggregation:

ids=[0,2,4]

df = (df[df['id'].isin(ids)].groupby('name', as_index=False)
                            .agg(I=('in','sum'), C=('cy','sum'))
                            .rename(columns={'name':'N'}))
print (df)
    N   I   C
0  MD  36  14
1  ZD  14  10

CodePudding user response：

You can use isin then groupby and named aggregation

out = df[df['id'].isin(ids)].groupby('name').agg(I=('in', sum),
                                                 C=('cy', sum)).reset_index().rename(columns={'name': 'N'})

print(out)

    N   I   C
0  MD  36  14
1  ZD  14  10