Function downsample 

pub fn downsample(
    batch: &RecordBatch,
    value_column: &str,
    bucket: &TimeBucket,
    function: &AggFn,
    memory_tracker: Option<&MemoryTracker>,
) -> Result<RecordBatch>

Downsample a RecordBatch into time buckets.

For each bucket [bucket.start_of(ts), bucket.next(start)), the aggregation function is applied to the values in the value_column. The bucket may be a fixed span or a calendar one — it is the same type SQL’s time_bucket() and a rollup tier use, so the three agree.

Returns a new RecordBatch with two columns: _time (bucket start) and the aggregated value column.

Arguments

• batch – input data (must have a _time Int64 column)
• value_column – name of the column to aggregate
• bucket – the bucket width and the calendar it is read against
• function – aggregation function to apply per bucket

Errors

Returns an error if columns are missing or have unexpected types.