ISSUE-1781: col encoding mapping & stats accumulator

[dantengsky]

I hereby agree to the terms of the CLA available at: https://databend.rs/policies/cla/

Summary

fixed issues:

• statistics accumulator (by using facilities of common data block, not parquet) #2193
• interagte with data accessor #2194
• enable proper compression/encoding algos for columns #1781

NOTE

• transient modification of data_blcok.rs has been reverted
• "no side effects"
  crates / components other than dal, query/datasource/table/fuse are not touched, except that lz4 in enabled for parquet2

Changelog

• Improvement
• Not for changelog (changelog entry is not required)

Related Issues

Fixes #1781
Fixes #2193
Fixes #2194

Test Plan

Unit Tests

Stateless Tests

[databend-bot]

Thanks for the contribution!
I have applied any labels matching special text in your PR Changelog.

Please review the labels and make any necessary changes.

[codecov-commenter]

Codecov Report

Merging #2159 (93e72b4) into master (fbc7ae7) will increase coverage by 0%.
The diff coverage is 51%.

@@           Coverage Diff           @@
##           master   #2159    +/-   ##
=======================================
  Coverage      67%     68%            
=======================================
  Files         637     643     +6     
  Lines       35849   35967   +118     
=======================================
+ Hits        24361   24599   +238     
+ Misses      11488   11368   -120     

Impacted Files	Coverage Δ
common/dal/src/data_accessor.rs	0% <ø> (ø)
common/dal/src/impls/aws_s3/s3.rs	0% <ø> (ø)
...uery/src/datasources/table/fuse/io/block_reader.rs	0% <0%> (ø)
query/src/datasources/table/fuse/io/mod.rs	0% <ø> (ø)
query/src/datasources/table/fuse/table.rs	0% <0%> (ø)
...uery/src/datasources/table/fuse/table_do_append.rs	0% <0%> (ø)
query/src/datasources/table/fuse/table_do_read.rs	0% <0%> (ø)
...y/src/datasources/table/fuse/table_do_read_plan.rs	0% <0%> (ø)
...ry/src/datasources/table/fuse/table_do_truncate.rs	0% <0%> (ø)
...y/src/datasources/table/fuse/util/index_helpers.rs	0% <ø> (ø)
... and 22 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fbc7ae7...93e72b4. Read the comment docs.

[BohuTANG]

For datablock, we convert it to RecordBatch, 'TryInto' is better?

[dantengsky]

worrying about the cost of conversion here?

seems the cost is the same whether converted from DataBlock or a reference of DataBlock. but extra pointers may still reference the underlying data though.

currently, some code needs to access DataBlock again after the conversion, which is being refactored.

later, this modification will be reverted.

[sundy-li]

So we can use TryFrom::try_from(&v) for impl TryFrom<DataBlock> for RecordBatch to reduce to code size.

[dantengsky]

good idea.

I am changing the caller side code so that Datablock will no longer need to be accessed after RecordBatch::try_from.

this is just a tmp modification of data_block.rs, to testify some statistic component, will not be submit in the final PR

[BohuTANG]

/lgtm

Great~

[databend-bot]

Approved! Thank you for the PR @dantengsky

[databend-bot]

CI Passed
Reviewer Approved
Let's Merge