// the find
x-ream/sqli
orm sql query builder, API: QB, QB.X, QrB
A Java SQL query builder plus a repository-pattern persistence layer (BaseRepository/RepositoryX) that bolts onto Spring Boot, with dialect support stretching from MySQL to analytical engines like ClickHouse, Impala and TDengine. Aimed at Java/Spring shops that want a fluent SQL layer instead of JPA/Hibernate's full ORM magic.
The query builder API (QB/QB.X/QrB) is genuinely expressive for hand-rolled SQL: conditional joins with multi-predicate ON clauses, sub-queries, GROUP BY/HAVING via a Reduce API, and raw-fragment escape hatches (x()) when the builder can't express something. The dialect module covering ClickHouse/Impala/TDengine alongside MySQL/Oracle is a real differentiator — most Java query builders only target standard RDBMS. Some pragmatic built-in optimizations: unused JOINs get stripped from generated SQL automatically, and IN clauses are chunked at 500 items to dodge driver/DB limits.
Documentation is almost entirely in Chinese with broken/garbled English in spots — a serious barrier for most of the potential audience, and the actual framework integration (x7-spring-boot-starter) lives in a separate, undocumented-here repo, so you're piecing together two projects to get started. The L2 cache feature is a footgun: it's a manual-discipline system where you must remember to call CacheFilter.filter() on every write path or silently get stale cached reads, and results over 20 rows quietly lose any caching benefit with no enforcement or warning in code. No UNION support at all (explicitly dropped as 'too complex'), which rules out a common real-world query pattern. The custom ID generator and proxy-based repository interfaces (no implementation class, just an annotated interface) mean debugging requires understanding the framework's bytecode/proxy internals rather than stepping through plain code.