New bloom filter index for text, sparse_gram

Why it matters

The sparse_gram bloom filter index is designed to efficiently speed up text search queries by providing a compact and fast probabilistic index. It improves query performance for text data by quickly filtering out non-matching rows with minimal memory usage.

How to use it

To use the sparse_gram bloom filter index, create an index on the targeted text column using the bloom_filter index type with the sparse_gram granularity parameter. For example:

CREATE TABLE example (
    text_column String,
    INDEX sparse_gram_index (text_column) TYPE bloom_filter('sparse_gram') GRANULARITY 1
) ENGINE = MergeTree()
ORDER BY tuple();