pandas pivot: Index contains duplicate entries, cannot reshape (tested)

ValueError: Index contains duplicate entries, cannot reshape from df.pivot() means some (index, columns) pair appears in more than one row. pivot cannot combine them; use pivot_table with an explicit aggfunc, or drop the duplicates first. Tested on pandas 2.3.3 and 3.1.0.

1. Reproduce it

import pandas as pd
df = pd.DataFrame({
    "date":  ["2026-01-01", "2026-01-01", "2026-01-01", "2026-01-02"],
    "store": ["A", "A", "B", "A"],
    "sales": [10, 5, 7, 3],
})
try:
    df.pivot(index="date", columns="store", values="sales")
except ValueError as e:
    print("ValueError:", e)

Output, identical on pandas 2.3.3 and 3.1.0:

ValueError: Index contains duplicate entries, cannot reshape

pivot() only reshapes; it never combines values. Each (index, columns) pair must appear once, because it becomes exactly one cell. Here ("2026-01-01", "A") appears twice (10 and 5), so pandas cannot decide which to put in the cell. The "Index" in the message means that pair, not your DataFrame index, so df.index.is_unique being True does not help.

2. Find the duplicates, then aggregate them

import pandas as pd
df = pd.DataFrame({
    "date":  ["2026-01-01", "2026-01-01", "2026-01-01", "2026-01-02"],
    "store": ["A", "A", "B", "A"],
    "sales": [10, 5, 7, 3],
})
# 1. find the rows that collide
dupes = df[df.duplicated(["date", "store"], keep=False)]
print(dupes)
# 2. aggregate them: pivot_table (default aggfunc is mean)
print(df.pivot_table(index="date", columns="store", values="sales", aggfunc="sum"))
# 3. same result with groupby + unstack
print(df.groupby(["date", "store"])["sales"].sum().unstack())

Output, identical on pandas 2.3.3 and 3.1.0:

         date store  sales
0  2026-01-01     A     10
1  2026-01-01     A      5
store          A    B
date                 
2026-01-01  15.0  7.0
2026-01-02   3.0  NaN
store          A    B
date                 
2026-01-01  15.0  7.0
2026-01-02   3.0  NaN

duplicated(..., keep=False) lists every colliding row so you can check whether they are real repeats (sum them) or bad data (drop them with drop_duplicates(["date", "store"]) and then pivot works). pivot_table with no aggfunc uses the mean: on this data the A cell becomes 7.5, not 15, which is easy to miss. Add fill_value=0 to get 0 instead of NaN for missing pairs; the columns then stay integers (15, 7, 3, 0).

3. Other pivot errors

import pandas as pd
df = pd.DataFrame({"date": ["d1", "d2"], "store": ["A", "B"], "sales": [1, 2]})
try:
    df.pivot(index="Date", columns="store", values="sales")
except KeyError as e:
    print("KeyError:", e)
try:
    df.pivot("date", "store", "sales")
except TypeError as e:
    print("TypeError:", e)

Output with pandas 2.3.3:

KeyError: 'Date'
TypeError: DataFrame.pivot() takes 1 positional argument but 4 were given

Output with pandas 3.1.0:

KeyError: "The following 'index' labels are not columns of the DataFrame: ['Date']"
TypeError: DataFrame.pivot() takes 1 positional argument but 4 were given

A misspelled column (here Date vs date) gives a bare KeyError on pandas 2; pandas 3 names the argument. Since pandas 2.0, index, columns and values must be passed as keywords; old tutorials that call df.pivot("date", "store", "sales") now raise TypeError.

Related: KeyError but the column exists · groupby FutureWarnings · merge KeyError / not matching.

Written with AI assistance. Measured on 2026-10-23 in the python:3-alpine Docker image (Python 3.14.7) with pandas 2.3.3 and pandas 3.1.0. Outputs pasted unedited. PlainTables is our extension; Python and pandas are free.