Categorical in dataframe is sorted lexically.

[has2k1]

code

df = pd.DataFrame({"id":[6,5,4,3,2,1], "raw_grade":['a', 'b', 'b', 'a', 'a', 'e']})
df["grade"] = pd.Categorical(df["raw_grade"])
df['grade'].cat.reorder_levels(['b', 'e', 'a'])

# sorts 'grade' according to the order of the levels
df.sort(columns=['grade'])  

correct output

  id   raw_grade  grade
4 5    b          b
3 4    b          b
0 1    e          e
2 6    a          a
1 3    a          a
5 2    a          a

---

code

# sorts 'grade' lexically
df.sort(columns=['grade', 'id'])

wrong output

  id   raw_grade  grade
4 2    a          a
3 3    a          a
0 6    a          a
2 4    b          b
1 5    b          b
5 1    e          e

When there is more than one element in the columns list, the Categoricals columns are sorted lexically.

pandas: 0.14.1-78-g24b309f

[jreback]

after #7850

In [1]: df = DataFrame({"id":[6,5,4,3,2,1], "raw_grade":['a', 'b', 'b', 'a', 'a', 'e']})

In [2]: df["grade"] = pd.Categorical(df["raw_grade"])

In [3]: df['grade'].cat.reorder_levels(['b', 'e', 'a'])

In [4]: df
Out[4]: 
   id raw_grade grade
0   6         a     a
1   5         b     b
2   4         b     b
3   3         a     a
4   2         a     a
5   1         e     e

In [5]: df.dtypes
Out[5]: 
id              int64
raw_grade      object
grade        category
dtype: object

In [6]: df.sort(columns=['grade'])
Out[6]: 
   id raw_grade grade
1   5         b     b
2   4         b     b
5   1         e     e
0   6         a     a
3   3         a     a
4   2         a     a

In [7]: df.sort(columns=['grade', 'id'])
Out[7]: 
   id raw_grade grade
2   4         b     b
1   5         b     b
5   1         e     e
4   2         a     a
3   3         a     a
0   6         a     a