microsoft/qlib · error · ValueError

lib {} not in arctic

Error message

lib {} not in arctic

What it means

The RemoteArcticProvider feature loader (qlib/contrib/data/data.py) reads data from an Arctic (MongoDB-backed) timeseries store, keyed by libraries named after frequency ('day', '1min', ...). If the Arctic instance has no library matching the requested freq, it raises ValueError('lib {} not in arctic'). The check is deliberately early so a missing dataset fails before symbol lookup.

Source

Thrown at qlib/contrib/data/data.py:39

        self, uri="127.0.0.1", retry_time=0, market_transaction_time_list=[("09:15", "11:30"), ("13:00", "15:00")]
    ):
        super().__init__()
        self.uri = uri
        # TODO:
        # retry connecting if error occurs
        # does it real matters?
        self.retry_time = retry_time
        # NOTE: this is especially important for TResample operator
        self.market_transaction_time_list = market_transaction_time_list

    def feature(self, instrument, field, start_index, end_index, freq):
        field = str(field)[1:]
        with pymongo.MongoClient(self.uri) as client:
            # TODO: this will result in frequently connecting the server and performance issue
            arctic = Arctic(client)

            if freq not in arctic.list_libraries():
                raise ValueError("lib {} not in arctic".format(freq))

            if instrument not in arctic[freq].list_symbols():
                # instruments does not exist
                return pd.Series()
            else:
                df = arctic[freq].read(instrument, columns=[field], chunk_range=(start_index, end_index))
                s = df[field]

                if not s.empty:
                    s = pd.concat(
                        [
                            s.between_time(time_tuple[0], time_tuple[1])
                            for time_tuple in self.market_transaction_time_list
                        ]
                    )
                return s

View on GitHub (pinned to 79633dd950)

Solutions

1. Verify what exists: arctic.list_libraries() against the same MongoDB URI, and request only those frequencies.
2. Create/populate the missing library with your data ingestion pipeline (initialize the Arctic lib for that freq and write instrument data).
3. Point provider_uri at the correct Mongo host/database where the library actually lives.
4. Catch the ValueError at startup and degrade to an available frequency rather than crashing mid-run.

Example fix

# before
D.features(instruments, fields, freq='1min')  # ValueError if '1min' lib absent
# after
from arctic import Arctic
libs = Arctic(self.uri).list_libraries()
freq = '1min' if '1min' in libs else 'day'
D.features(instruments, fields, freq=freq)

from arctic import Arctic
from pymongo import MongoClient
with MongoClient(uri) as mc:
    libs = Arctic(mc).list_libraries()
if freq not in libs:
    raise ValueError(f'freq {freq} not in arctic; available: {libs}')

try:
    feats = D.features(insts, fields, freq=freq)
except ValueError as e:
    if 'not in arctic' in str(e):
        feats = D.features(insts, fields, freq='day')  # degrade to available freq
    else:
        raise

When it happens

Trigger: Configuring qlib with an arctic provider_uri ('arctic_host:port/lib') and requesting a frequency whose library was never created — e.g. requesting 1min data when only the 'day' library was ever initialized on the Arctic/MongoDB server.

Common situations: Production MongoDB migrations that rename or drop Arctic libraries; using a fresh Mongo instance without running the data-dump scripts; requesting 'tick'/'1min' freq against a day-only store.

Related errors

• Invalid mount path
• Unknown mount error: {error_output.strip()}
• Failed to create directory {mount_path}, please create {moun
• nfs-common is not found, please install it by execute: sudo
• Mount failed: requires sudo or permission denied

AI-assisted analysis of microsoft/qlib@79633dd950 (2026-08-15). Data as JSON: /api/errors/88b0a62b773d751c. Report an issue: GitHub.