pola-rs/polars · error · TypeError

`pl.count()` takes at least one argument. If you want to cou

Error message

`pl.count()` takes at least one argument. If you want to count the number of rows, please use `pl.len()` instead.

What it means

TypeError raised by pl.count() when called with no column arguments. pl.count() without columns historically returned the row count, which was ambiguous and is now reserved for counting non-null values in specified columns; the message directs users to pl.len() for row counts.

Source

Thrown at py-polars/src/polars/functions/lazy.py:224

    │ 1   ┆ 3   │
    └─────┴─────┘

    Return the number of rows in a context. **This way of using the function is
    deprecated.** Please use :func:`len` instead.

    >>> df.select(pl.count())  # doctest: +SKIP
    shape: (1, 1)
    ┌───────┐
    │ count │
    │ ---   │
    │ u32   │
    ╞═══════╡
    │ 3     │
    └───────┘
    """
    if not columns:
        msg = "`pl.count()` takes at least one argument. If you want to count the number of rows, please use `pl.len()` instead."
        raise TypeError(msg)
    return F.col(*columns).count()


def cum_count(*columns: str, reverse: bool = False) -> Expr:
    """
    Return the cumulative count of the non-null values in the column.

    This function is syntactic sugar for `col(columns).cum_count()`.

    Parameters
    ----------
    *columns
        Name(s) of the columns to use.
    reverse
        Reverse the operation.

    Examples
    --------

View on GitHub (pinned to 68506541d2)

Solutions

Replace pl.count() with pl.len() for row counts
Use pl.count('col') to count non-null values of a specific column
Grep for pl.count() during upgrades — the fix is mechanical

Example fix

# before
df.group_by('a').agg(pl.count())
# after
df.group_by('a').agg(pl.len())

Defensive patterns

Strategy: validation

Validate before calling

if not columns:
    raise TypeError('use pl.len() for row counts')

Prevention

Prefer pl.len() everywhere for row counts
Reserve pl.count('col') for non-null counts
Grep and replace bare pl.count() on upgrades

When it happens

Trigger: pl.count() with zero arguments, e.g. df.group_by('a').agg(pl.count()) or pl.select(pl.count()) on older-style code.

Common situations: Tutorial/legacy code predating the pl.len() introduction; SQL-transliteration habits where COUNT(*) means row count; group_by aggregations copied from old docs.

Related errors

AI-assisted analysis of pola-rs/polars@68506541d2 (2026-08-28). Data as JSON: /api/errors/95b60e052fcfdae7. Report an issue: GitHub.