Pandas switching value-CodePudding

I am trying to switch the group number by the people of number, the table like this below

|   uid  |  group |
| ------ | ------ |
| 355    | 1      |
| 356    | 3      |
| 357    | 4      |
| 358    | 3      |
| 359    | 5      |
| 360    | 6      |...

want to switch the number if the people in group 1 rename to group 5, if in group 3 rename to group 6... how can I able to change the value? I try to

df['group'] = df['group'][df['group'] == 1]   4 
df['group'] = df['group'][df['group'] == 3]   3

but i also calculate the NAN value to the dataframe i also try to

df['group'][df['group'] == 5] = df['group'][df['group'] == 10]

by add all group by 10 from the start, but not work

CodePudding user response：

Use a mapping dictionary:

d = {1: 5, 3: 6}

df['group'] = df['group'].map(lambda x: d.get(x, x))

Or if you want to use an addition:

d = {1: 4, 3: 3}

df['group'] = df['group'].add(df['group'].map(d), fill_value=0)

output (as new column "group2" for clarity):

   uid  group  group2
0  355      1       5
1  356      3       6
2  357      4       4
3  358      3       6
4  359      5       5
5  360      6       6

CodePudding user response：

you can use loc for this and assign the group with it

df.loc[ {your condition}, {the column that you want to update}] = {new value}

by replacing those {}, for your first example this is the code

df = pd.DataFrame({
    'uid': [355,356,357,358,359,360],
    'group': [1, 2, 3, 4, 5,6]
})

df.loc[df["group"]==1,"group"]=5
df.loc[df["group"]==3,"group"]=6

the output is this