How to Reset Index in pandas (DataFrame, groupby, drop=True) [Examples]

If your row labels no longer line up with the table—common after a groupby, a filter, or a sort—reset_index() is how you put the index back in order. By default it turns the current index into a normal column and gives you a fresh 0, 1, 2, … down the side; with drop=True it skips that extra column and only keeps the new numbering. The examples below use both DataFrame and Series.

Tested environment: Every python block below was run and tested with python3 on Ubuntu (Python 3.13.3, pandas 2.2.3).

Quick reference

After reshaping data, you can print the entire DataFrame to confirm row labels and columns.

Default reset_index() on a DataFrame

By default, reset_index() moves the row index into a new column and assigns a new integer index from 0 upward—useful when strings, dates, or ids on the index should become ordinary columns again.

import pandas as pd

df = pd.DataFrame({"A": [1, 2, 3]}, index=["x", "y", "z"])
print(df)
print("---")
df_reset = df.reset_index()
print(df_reset)

A
x  1
y  2
z  3
---
  index  A
0     x  1
1     y  2
2     z  3

The new column is named index unless your index had a name.

drop=True: reset index to start at 0 without an extra column

For pandas reset index to start at 0 and no added column, pass drop=True. Use this when the old labels are not useful anymore (for example after sorting or subsetting).

import pandas as pd

df = pd.DataFrame({"A": [1, 2]}, index=["x", "y"])
print(df.reset_index(drop=True))

A
0  1
1  2

Pandas groupby reset index

After groupby, group keys often sit on the index. Calling reset_index() turns them back into regular columns — the usual pandas groupby reset index pattern. For a deeper tour of grouping, see the pandas groupby guide on this site.

import pandas as pd

df = pd.DataFrame(
    {
        "A": [1, 2, 3, 1, 2, 3],
        "B": [4, 5, 6, 4, 5, 6],
        "C": [7, 8, 9, 7, 8, 9],
    }
)
grouped_df = df.groupby("A").sum().reset_index()
print(grouped_df)

A   B   C
0  1   8  14
1  2  10  16
2  3  12  18

Alternative (no reset_index call): df.groupby("A", as_index=False).sum() produces the same aggregated values for this data; choose it when you never need the grouped object as a MultiIndex and want to skip an extra step.

MultiIndex rows

If the row index has multiple levels, reset_index() flattens each level into its own column unless you pass level.

import pandas as pd

arrays = [["a", "a", "b", "b"], [0, 1, 0, 1]]
idx = pd.MultiIndex.from_arrays(arrays, names=["letter", "num"])
df = pd.DataFrame({"value": [10, 20, 30, 40]}, index=idx)
print(df.reset_index())

letter  num  value
0      a    0     10
1      a    1     20
2      b    0     30
3      b    1     40

Rename the new column after reset_index()

If you keep the old index as a column, you may want a clearer name than index.

import pandas as pd

df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6], "C": [7, 8, 9]}, index=["a", "b", "c"])
df = df.reset_index().rename(columns={"index": "original_index"})
print(df)

original_index  A  B  C
0              a  1  4  7
1              b  2  5  8
2              c  3  6  9

Start row labels at 1 instead of 0

The default index after reset_index() starts at 0. To start at 1, reset first then shift the RangeIndex (avoid off-by-one on empty frames in real pipelines).

import pandas as pd

df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6], "C": [7, 8, 9]}, index=["a", "b", "c"])
df = df.reset_index()
df.index = df.index + 1
print(df)

index  A  B  C
1     a  1  4  7
2     b  2  5  8
3     c  3  6  9

Series reset_index()

Series.reset_index() returns a DataFrame with two columns: the old index (name index by default) and the series values (often column 0 until you rename).

import pandas as pd

s = pd.Series([1, 2, 3], index=["a", "b", "c"])
print(s)
print("---")
print(s.reset_index())

a    1
b    2
c    3
dtype: int64
---
  index  0
0     a  1
1     b  2
2     c  3

After sort_values()

Sorting keeps the original index attached to each row. Reset index in pandas after sort when you want row numbers to follow the new row order.

import pandas as pd

df = pd.DataFrame({"A": [3, 2, 1], "B": [6, 5, 4], "C": [9, 8, 7]}, index=["c", "b", "a"])
sorted_df = df.sort_values("A").reset_index()
print(sorted_df)

index  A  B  C
0     a  1  4  7
1     b  2  5  8
2     c  3  6  9

After query() or boolean filtering

Filtered rows keep old index labels, so you see gaps in the index sequence. reset_index() fixes reset index in pandas after subsetting.

import pandas as pd

df = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6], "C": [7, 8, 9]}, index=["a", "b", "c"])
filtered_df = df.query("A > 1").reset_index()
print(filtered_df)

index  A  B  C
0     b  2  5  8
1     c  3  6  9

Pandas reset column index (column labels 0, 1, …)

Row reset_index() does not rename columns. For pandas reset column index, assign integer column labels after you are done shaping the frame:

import pandas as pd

df = pd.DataFrame([[10, 20], [30, 40]], columns=["foo", "bar"])
print("before:")
print(df)
df.columns = range(df.shape[1])
print("after:")
print(df)

before:
   foo  bar
0   10   20
1   30   40
after:
    0   1
0  10  20
1  30  40

Syntax and parameters (reference)

From the official API, the signature is essentially:

DataFrame.reset_index(level=None, drop=False, inplace=False, col_level=0, col_fill='', ...)

• level — reset only some levels of a MultiIndex.
• drop — if True, discard the old index instead of adding a column.
• inplace — if True, mutate df and return None.
• col_level / col_fill — placement when columns are MultiIndex.

Newer pandas versions add extra keyword arguments; see pandas.DataFrame.reset_index for your exact version.

To put one or more columns onto the index instead, use set_index — see pandas set index on this site.

Summary

Reset index pandas workflows boil down to: use reset_index() when the index should become data again, reset_index(drop=True) when you only want a clean 0..n-1 dataframe reset index, and groupby(...).sum().reset_index() (or as_index=False) for groupby reset index results. For column labels, assign df.columns. All examples above were verified on pandas 2.2.3 so you can trust the printed output.

References

• pandas.DataFrame.reset_index (official)