Changelog category (leave one):

• Performance Improvement

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

Add optional .size subcolumn serialization for top-level String columns in MergeTree tables to improve compression and enable efficient subcolumn access. Introduce new MergeTree settings for serialization version control and expression optimization for empty strings.

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

• New MergeTree settings:

  • serialization_info_version – Controls serialization info format when writing serialization.json. Required for cluster upgrades.

    • DEFAULT – Legacy format, compatible with old servers during rolling upgrades.
    • WITH_TYPES – New format with types_serialization_versions, enabling per-type serialization settings like string_serialization_version. Switch to this after upgrades.

  • string_serialization_version – Controls top-level String column serialization (effective only when serialization_info_version = WITH_TYPES).

    • DEFAULT – Standard inline size format.
    • WITH_SIZE_STREAM – Serialize top-level String columns with separate .size stream for better compression. Backward incompatible.

• Subcolumn support:

  • Access .size subcolumn across both legacy and new String formats, supporting mixed-format queries.

• Expression optimizations:

  • optimize_empty_string_comparisons rewrites str = '' into isEmpty(str)/isNotEmpty(str).
  • FunctionToSubcolumnsPass extended to rewrite length(str) as str.size.

• Sparse encoding enhancements:

  • Support for Sparse encoding with multiple substreams for future extension.