microsoft/qlib · error · ValueError

can't find a freq from {self.support_freq} that can resample

Error message

can't find a freq from {self.support_freq} that can resample to {self.freq}!

What it means

Raised by FileCalendarStorage._freq_file (qlib/data/storage/file_storage.py) when the requested calendar freq is not directly available and no stored freq can serve as a resampling source. Qlib can resample calendars from a finer available freq (via Freq.get_recent_freq); if none of support_freq can be resampled into the target freq, this ValueError is raised. Typical when requesting a coarser or unrelated freq than everything on disk.

Source

Thrown at qlib/data/storage/file_storage.py:101

    @property
    def file_name(self) -> str:
        return f"{self._freq_file}_future.txt" if self.future else f"{self._freq_file}.txt".lower()

    @property
    def _freq_file(self) -> str:
        """the freq to read from file"""
        if not hasattr(self, "_freq_file_cache"):
            freq = Freq(self.freq)
            if freq not in self.support_freq:
                # NOTE: uri
                #   1. If `uri` does not exist
                #       - Get the `min_uri` of the closest `freq` under the same "directory" as the `uri`
                #       - Read data from `min_uri` and resample to `freq`

                freq = Freq.get_recent_freq(freq, self.support_freq)
                if freq is None:
                    raise ValueError(f"can't find a freq from {self.support_freq} that can resample to {self.freq}!")
            self._freq_file_cache = freq
        return self._freq_file_cache

    def _read_calendar(self) -> List[CalVT]:
        # NOTE:
        # if we want to accelerate partial reading calendar
        # we can add parameters like `skip_rows: int = 0, n_rows: int = None` to the interface.
        # Currently, it is not supported for the txt-based calendar

        if not self.uri.exists():
            self._write_calendar(values=[])

        with self.uri.open("r") as fp:
            res = []
            for line in fp.readlines():
                line = line.strip()
                if len(line) > 0:
                    res.append(line)

View on GitHub (pinned to 79633dd950)

Solutions

Request one of the freqs actually stored (check calendars/ folder stems) — resampling from a stored freq is only attempted, not guaranteed to succeed.
Provide a finer-grained calendar/calendar data on disk (e.g. dump 1min calendars) so the target freq can be resampled from it.
Verify freq strings use qlib's canonical units ('day', '1min', '5min', ...); unsupported spellings can never match or resample.
As a workaround, build week/month results yourself by resampling the daily loaded dataframe with pandas after fetching at 'day'.

Example fix

# before
cal = Cal.calendar(freq="week")  # only calendars/day.txt on disk -> ValueError

# after
cal = Cal.calendar(freq="day")
# or dump a finer calendar (e.g. 1min) so 'week' can be resampled from it

Defensive patterns

Strategy: validation

Validate before calling

from qlib.data.storage.file_storage import FileCalendarStorage
try:
    FileCalendarStorage(freq, future).uri
    ok = True
except ValueError:
    ok = False

Try / catch

try:
    cal = Cal.calendar(freq=freq)
except ValueError as e:
    if "can't find a freq" in str(e):
        cal = Cal.calendar(freq="day")  # fall back to stored freq, resample yourself
    else:
        raise

Prevention

Prefer requesting freqs that exist verbatim as calendar .txt files.
If you need week/month bars, fetch daily and resample with pandas in your pipeline.

When it happens

Trigger: Only a 'day' calendar exists on disk and you request freq='week' or 'month' without a finer source available for resampling logic; requesting '5min' when only 'day' exists (cannot upsample from coarse to fine); freq taxonomies that do not relate (e.g. 'tick' vs 'day').

Common situations: Asking qlib to derive weekly/monthly calendars without supplying the underlying daily/minute data; mismatched freq naming conventions between the request and dumped files; partial data downloads missing the fine-grained calendars.

Related errors

AI-assisted analysis of microsoft/qlib@79633dd950 (2026-08-15). Data as JSON: /api/errors/fe5b15fa1108e234. Report an issue: GitHub.