Use case

To check how the data compresses without writing it.

Describe the solution you'd like

A parametric aggregate function

estimateCompressionRatio('codec', block_size_bytes)(column)

Both parameters are optional.

The function accepts a single column. It will serialize it using binary serialization and put it into a compressing buffer of the specified size on top of a Null destination buffer. So, the buffer is compressed, and the results are discarded, but the size is counted.

It returns the ratio between the uncompressed and compressed data (e.g., 2 means the data is compressed twice).

Describe alternatives you've considered

I thought we could allow multiple columns and use the Native format. However, it will require constructing temporary columns and using temporary blocks in memory before serialization.