Sorting values after transpose of dataframe-CodePudding

My dataframe is:

0	1	2	3
a	1091	347	2164
b	208	284	27647
c	0	8126	22

After transposing

0	a	b	c
1	1.091	208	0
2	347	284	8126
3	2161	27647	22

df=df.set_index(0)
print(df)
df = df.T
print(df)
df = df.sort_values(by=df.index, ascending=True)
print(df)

But I am getting an error like:

KeyError: Int64Index([1, 2, 3, 4], dtype='int64')

I am trying to plot index values to a column but it should be in ascending order because index 3 refers new value but index 1 is an old value in the time range.

fig = px.line(df, x=df.index, y=df['a'])
fig.show()

CodePudding user response：

Instead of df.sort_values(by=df.index), use df.sort_index:

df = df.sort_index(ascending=True)

Output:

>>> df
      1     2      3
a  1091   347   2164
b   208   284  27647
c     0  8126     22

CodePudding user response：

You can sort the index if you cast it to str dtype with a list comprehension.

df = pd.DataFrame({'0': {0: 'a', 1: 'b', 2: 'c'},
 '1': {0: 1091, 1: 208, 2: 0},
 '2': {0: 347, 1: 284, 2: 8126},
 '3': {0: 2164, 1: 27647, 2: 22}})

df = df.T
df=df.set_index(0)
df.index = [str(x) for x in df.index]
df.sort_index(ascending=True)

Output df:

        1       2
1091    208 0
2164    27647   22
347     284     8126
a       b       c

CodePudding user response：

i would prefer to sort by specific column. sort_index sorts by first column afaik.

df.sort_values('column name')