Suppose I have a dataframe

df =
| CHROM | 
| ------| 
| chr1  | 
| chr5  |
| chr12 |
| chr9  |
| chr3  |

I have used this code to sort them:

sorted_df = df.sort_values(by=["CHROM"])

I got the result,

| CHROM | 
| ------| 
| chr1  | 
| chr12 |
| chr3  |
| chr5  |
| chr9  |

But, My expected output

| CHROM | 
| ------| 
| chr1  | 
| chr3  |
| chr5  |
| chr9  |
| chr12 |

Please suggest how to do using python

CodePudding user response：

You can split base "chr" then sort the base number as int and find the index and return df with the sorted index.

idx = df["CHROM"].str.split('chr').str[-1].astype(int).sort_values().index
df_new = df.iloc[idx]
print(df_new)

Or You can use pandas.DataFrame.sort_values with key.

df.sort_values(by=['CHROM'], 
               key=lambda x: df['CHROM'].str.split('chr').str[-1].astype(int))

Output:

   CHROM
0   chr1
4   chr3
1   chr5
3   chr9
2  chr12