Query parameter combinations
This guide explains the main parameters for CloudQuant Data Liberator queries and the expected results from different combinations.
Core parameters
| Parameter | Description | Default |
|---|
name | Dataset identifier (required) | — |
symbols | Single symbol or list, e.g. ['AAPL','GOOG','MSFT'] | None (all symbols) |
back_to | Query start date | None |
as_of | Query end date | Current date/time |
fields | Specify desired columns to reduce download time | All columns |
stats | Set to 'Total' for symbol counts instead of full results | — |
where | Custom SQL predicate appended to the generated filters | None |
sql | Full raw SQL query (replaces name, time range, and other construction params) | None |
Use liberator.datasets(entitled=True) to list all datasets available to you.
Query examples
All parameters included
df = liberator.get_dataframe(liberator.query(
name = 'daily_bars',
as_of = '2024-07-24',
back_to = '2024-07-16',
symbols = ['AAPL', 'GOOGL']))
Without symbols
df = liberator.get_dataframe(liberator.query(
name = 'daily_bars',
as_of = '2024-07-24',
back_to = '2024-07-16'))
Without as_of
df = liberator.get_dataframe(liberator.query(
name = 'daily_bars',
back_to = '2024-07-22'))
as_of to current time, returning data from present back to specified date.
Without back_to
df = liberator.get_dataframe(liberator.query(
name = 'daily_bars',
as_of = '2023-12-25'))
Minimal query (name only)
df = liberator.get_dataframe(liberator.query(name = 'daily_bars'))
Custom where filters
Use where with a standard parameter-based query to append extra SQL predicates on top of the filters CloudQuant Data Liberator generates for time range and symbols. The clause is combined with AND — do not include the WHERE keyword.
df = liberator.get_dataframe(liberator.query(
name='daily_bars',
as_of='2024-07-24',
back_to='2024-07-16',
symbols=['AAPL', 'GOOGL'],
where='volume > 1000000'
))
where works with time-range, LKV, and N-Query modes. Column names that match the dataset schema are automatically quoted when needed.
Multi-dataset / concordance-style filters
For datasets keyed by more than symbol (for example concordance tables with a dataset column), use where to narrow the result set:
df = liberator.get_dataframe(liberator.query(
name='concordance',
as_of='2025-01-01',
back_to='2024-01-01',
where="dataset = 'my_derived_dataset'"
))
where cannot be combined with sql. Use one approach or the other.
Raw sql queries
Pass a complete SQL SELECT statement via sql instead of name, back_to, as_of, and related parameters. When sql is present, parameter-based query construction is skipped.
df = liberator.get_dataframe(liberator.query(
sql='''
SELECT *
FROM "daily_bars"
WHERE muts >= 1704067200000000
AND muts < 1704153600000000
AND symbol = $$AAPL$$
''',
user='my_user',
system='API'
))
SQL requirements
- The statement must be a
SELECT (including UNION / INTERSECT / EXCEPT branches).
- Every
SELECT arm must include an explicit WHERE clause.
- By default, the
WHERE clause must include a time-range predicate on muts or another configured column (=, >, >=, <, <=, or BETWEEN). Use microsecond timestamps for muts values.
- Quote dataset names with double quotes when they contain special characters.
- Use
$$symbol$$ dollar-quoting for string literals in filters.
Parameters allowed with sql
When using sql, only connection, auth, and transfer options are permitted alongside the statement — for example user, system, token, compress, json_xfer, and batch_size. Supplying non-empty values for name, back_to, as_of, where, or other construction parameters raises an error.
Raw SQL bypasses the convenience of name / symbols / back_to / as_of. Prefer parameter-based queries unless you need full control over the generated SQL.
