[pandas] Sort

Basically, in pandas, there are two basic ways to sort - (1) sort by index and (2) sort by value.

 

1) Sort by index

df = pd.DataFrame({'a': [2,3,2,7,4], 'b': [2,1,3,5,3], 'c': [1,1,2,3,5]})

# ascending
print(df.sort_index())

# descending
print(df.sort_index(acending=False))

If you want to reset the index information as the one we sorted,

df.sort_index(ascending=False, inplace=True)
# (1)
print(df.reset_index())

# (2)
print(df.reset_index(drop=True))

(1) You can see we created a new index information. However, original index information still exists

(2) When you drop=True, then the original index information is completely replaced by newly created index information.

 

2) Sort by value

df = pd.DataFrame({'a': [2,3,2,7,4], 'b': [2,1,3,5,3], 'c': [1,1,2,3,5]})

# sort by value - standard: a specific column e.g., by a column 'a'
df.sort_values(by=['a'], inplace=True)
print(df)

# descending
print(df.sort_values(by=['a'], ascending=False, inplace=True))
print(df)

# sort by value - standard: several columns
df.sort_values(by=['a', 'b'], inplace=True)
print(df)

If we sort by value as a standard of several columns, we might misunderstand that both column 'a' and 'b' are ascendingly sorted such as 'a': [2,2,3,4,7] and 'b': [1,2,3,3,5].

However, code is written as "...by=['a', 'b']..." The program sorts by values according to 'a'. When there are same value in 'a' (e.g., 2 in 'a'), then within those value 'b' will be sorted.

Here is the result.

As we sorted the value, it is recommended to reset the index information.

df.reset_index(drop=True, inplace=True)
print(df)