redis/redis-py · error · ValueError

collect fields must be '*' or a non-empty list of names

Error message

collect fields must be '*' or a non-empty list of names

What it means

Raised by the collect reducer constructor when the fields argument is neither the literal '*' nor a non-empty collection of field names. The COLLECT reducer must project at least one field (or all fields via '*'); an empty selection is rejected locally before reaching the server.

Solutions

  1. Use collect() or collect(fields='*') to project all fields.
  2. Pass a non-empty list of real field names: collect(fields=['price', 'name']).
  3. Guard dynamic field lists: if not fields: fields = '*'.

Example fix

// before
r = collect(fields=field_list)  # field_list == []
// after
r = collect(fields=field_list or '*')
# or
r = collect(fields=['price'])
Defensive patterns

Strategy: validation

Validate before calling

fields = fields if (fields == '*' or (fields and all(str(f).strip() for f in fields))) else '*'
collect(fields=fields)

Try / catch

try:
    r = collect(fields=fields)
except ValueError:
    r = collect(fields='*')

Prevention

When it happens

Trigger: Calling collect(fields=[]), collect(fields=['']), collect(fields=[' ']), or collect(fields=()) — any iterable that yields no usable name.

Common situations: Dynamically building the field list from config/user input that resolved to empty. Whitespace-only field strings. Forgetting the '*' default.

Related errors


AI-assisted analysis of redis/redis-py@6a6b581b48 (2026-08-10). Data as JSON: /api/errors/bd11357e53e16767. Report an issue: GitHub.

Appendix: source

Thrown at redis/commands/search/reducers.py:236

        - **distinct**: When ``True``, emit ``DISTINCT`` to deduplicate entries
            with identical projected fields. Forward-compatible: this option is
            not yet implemented by the server and currently produces a server
            error when sent.
        - **sort_by**: An ``Asc``/``Desc`` instance or an iterable of them, used
            to order the collected entries within each group.
        - **limit**: An ``(offset, count)`` pair. Returns at most ``count``
            entries per group after skipping ``offset``. With ``sort_by`` this
            acts as a top-N selection.
        """
        args: list[str] = []

        # FIELDS (required)
        if fields == "*":
            args += ["FIELDS", "*"]
        else:
            names = [fields] if isinstance(fields, str) else list(fields)
            if not names or any(not n.strip() for n in names):
                raise ValueError(
                    "collect fields must be '*' or a non-empty list of names"
                )
            names = [_ensure_at_prefix(n) for n in names]
            args += ["FIELDS", str(len(names))] + names

        # DISTINCT (optional)
        if distinct:
            args += ["DISTINCT"]

        # SORTBY (optional)
        if sort_by is not None:
            sort_fields = [sort_by] if isinstance(sort_by, (Asc, Desc)) else sort_by
            sort_args: list[str] = []
            for f in sort_fields:
                sort_args += [_ensure_at_prefix(f.field), f.DIRSTRING]
            if not sort_args:
                raise ValueError("collect sort_by must contain at least one field")
            args += ["SORTBY", str(len(sort_args))] + sort_args

View on GitHub (pinned to 6a6b581b48)