--- title: Dolt squashes history 10x faster, Debezium 3.7 adds... description: Dolt's new dolt_squash_history() procedure collapses branch history in constant time, dropping a 100-commit squash from 1.6s to 0.15s and shrinking disk from... canonical: https://daily.dev/posts/dolt-squashes-history-10x-faster-debezium-3-7-adds-cockroachdb-sinkless-cdc-zkmhohqzx twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Dolt squashes history 10x faster, Debezium 3.7 adds CockroachDB sinkless CDC | daily.dev og:description: Dolt's new dolt_squash_history() procedure collapses branch history in constant time, dropping a 100-commit squash from 1.6s to 0.15s and shrinking disk from... og:url: https://daily.dev/posts/dolt-squashes-history-10x-faster-debezium-3-7-adds-cockroachdb-sinkless-cdc-zkmhohqzx og:image: https://api.daily.dev/og/posts/zKmHohqzX.png og:image:alt: Dolt squashes history 10x faster, Debezium 3.7 adds CockroachDB sinkless CDC og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Dolt squashes history 10x faster, Debezium 3.7 adds CockroachDB sinkless CDC **[Databases Digest](https://daily.dev/sources/databases_digest)** · 4 min read · 0 upvotes · 0 comments ## Summary Dolt's new dolt_squash_history() procedure collapses branch history in constant time, dropping a 100-commit squash from 1.6s to 0.15s and shrinking disk from 8.5MB to 172KB. Debezium 3.7.0.Alpha1 lands with 97 fixes, new Ingres and YashanDB connectors, and CockroachDB sinkless changefeed support that cuts out the Kafka middleman. DuckDB v2.0's async I/O shows a 20x speedup on CSV over S3 and 3x on Parquet. PostgreSQL 19's SQL/PGQ graph queries are real and useful for data lineage, but variable-length path support is missing and the syntax will put most people off. ## Content **TLDR:** Dolt's new dolt_squash_history() procedure collapses branch history in constant time, dropping a 100-commit squash from 1.6s to 0.15s and shrinking disk from 8.5MB to 172KB. Debezium 3.7.0.Alpha1 lands with 97 fixes, new Ingres and YashanDB connectors, and CockroachDB sinkless changefeed support that cuts out the Kafka middleman. DuckDB v2.0's async I/O shows a 20x speedup on CSV over S3 and 3x on Parquet. PostgreSQL 19's SQL/PGQ graph queries are real and useful for data lineage, but variable-length path support is missing and the syntax will put most people off. --- ## Dolt introduces dolt_squash_history(): 10x faster than rebase, 50x disk reduction after gc The new SQL procedure collapses an entire branch's commit history into a single commit using a pointer trick rather than replaying commits. A 100-commit squash drops from 1.6s with dolt_rebase() to 0.15s, and running dolt gc --full afterward shrinks disk from 8.5MB to 172KB. Partial squashing is supported via --first. History is rewritten, so treat remotes as archives before running this on anything shared. [Read more](https://daily.dev/posts/gbwjiEmTt) ## Debezium 3.7.0.Alpha1: CockroachDB sinkless CDC, Ingres connector, ExtendedTimestamp type The first 3.7 preview resolves 97 issues and ships incubating connectors for Ingres and YashanDB. The headline for CockroachDB users is sinkless changefeed support, which eliminates the need for an intermediate Kafka cluster. PostgreSQL gets first-class geometric type mapping, and the new ExtendedTimestamp type handles high-precision timestamps that overflow INT64. Breaking change: MySQL-to-MySQL JDBC sink pipelines are affected by type preservation fixes. [Read more](https://daily.dev/posts/61rTnGMjj) ## DuckDB v2.0 async I/O: 20x faster CSV over S3, 3x faster Parquet Scheduled for fall 2026, DuckDB v2.0 introduces async I/O using two thread pools — REGULAR workers for compute and ASYNC threads (defaulting to 4x system threads, capped at 256) for blocking I/O. TPC-H Q6 at SF100 shows Parquet queries on S3 going from 8.2s to 2.8s, and CSV from 878s to 45s. Concurrent query workloads jump from ~6 busy cores to ~48 out of 64. Read-ahead depth is tunable via read_ahead_depth, with memory pressure automatically throttling prefetch. [Read more](https://daily.dev/posts/p1OBh9KxZ) ## SQL/PGQ in PostgreSQL 19: useful for data lineage, but variable-length paths still need recursive CTEs PostgreSQL 19 ships SQL Property Graph Queries per ISO/IEC 9075-16:2023, letting you declare property graphs and query them with GRAPH_TABLE instead of chaining joins. The strongest use case is data lineage: model pipeline relationships as edge tables, declare a graph, and trace any aggregated metric back to source events in plain SQL. The honest limitation is that variable-length path queries aren't supported yet, so deep traversals still fall back to recursive CTEs. If your data is genuinely graph-shaped, this is worth learning; if you're reaching for it because it's new, you'll probably end up back at JOINs. [Read more](https://daily.dev/posts/0IynuoHLk) --- ## Also notable - **Percona Server for MySQL: stored procedure cursor metadata leaks RSS from 80GB to 180GB over 24 hours:** Cursor metadata allocated in MEM_ROOT during stored procedure calls is never freed for long-lived connections; under TPC-C load the RSS grew from ~80GB to over 180GB before triggering OOM kills, and the root cause exists in both MySQL 8.4 and 9.x — mitigations are limiting connection lifetime and adding brief pauses between transactions. [Read more](https://daily.dev/posts/7OukfMFhe) - **Airflow 3.0 Data Assets + deferrable operators cut worker CPU from 26% to 7.7% across 160 DAGs:** Halodoc migrated 30 DAGs to event-driven Data Assets scheduling and 130 to the deferrable RedshiftDataOperator, dropping base worker CPU from 26.1% to 7.71%, memory from 49.2% to 30.8%, and table-locking errors by ~38% — but an N+1 query bug in 3.0.6 pinned the metadata DB at 100% CPU, requiring an upgrade to 3.2.1. [Read more](https://daily.dev/posts/DaARqlW33) - **clickhousectl v0.4.0: horizontal autoscaling, ClickPipe schema discovery, MySQL CDC server-id option:** The v0.4.0 release adds horizontal autoscaling for ClickHouse Cloud services via CLI, beta ClickPipe schema discovery for previewing Kafka/Kinesis fields before pipe creation, and a --server-id flag for MySQL CDC pipes to avoid replication conflicts. [Read more](https://daily.dev/posts/Z6rj96vFI) - **Netflix device capability model: cumulative table plus 28-day histogram powers feature rollout decisions:** Netflix tracks hardware capabilities per device in a cumulative table and aggregates active device counts in a 28-day histogram broken down by model and software version, enabling distribution queries like "100% of devices support HD, 20% support UHD" that directly drive 4K and Spatial Audio rollout decisions. [Read more](https://daily.dev/posts/fGQdPlK1a) - **pgx CollectRows performance improvements land in pgx 5.6, covered in Go Weekly #612:** Go Weekly Issue #612 includes a deep dive into pgx 5.6's CollectRows performance improvements alongside a Go 1.28 proposal to add generic collection types including set.Set, ordered maps, and a generic binary heap. [Read more](https://daily.dev/posts/nQH8Z57m0) ## Similar posts on daily.dev - [Debezium 3.7.0.Beta1 Released](https://daily.dev/posts/debezium-3-7-0-beta1-released-r3h2ycqun) · Debezium · 5 upvotes · 1 comments --- Tags: [#backend](https://daily.dev/tags/backend), [#postgresql](https://daily.dev/tags/postgresql), [#duckdb](https://daily.dev/tags/duckdb), [#change-data-capture](https://daily.dev/tags/change-data-capture) [View this post on daily.dev](https://daily.dev/posts/dolt-squashes-history-10x-faster-debezium-3-7-adds-cockroachdb-sinkless-cdc-zkmhohqzx) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/dolt-squashes-history-10x-faster-debezium-3-7-adds-cockroachdb-sinkless-cdc-zkmhohqzx","headline":"Dolt squashes history 10x faster, Debezium 3.7 adds CockroachDB sinkless CDC","text":"Dolt's new dolt_squash_history() procedure collapses branch history in constant time, dropping a 100-commit squash from 1.6s to 0.15s and shrinking disk from 8.5MB to 172KB. Debezium 3.7.0.Alpha1 lands with 97 fixes, new Ingres and YashanDB connectors, and CockroachDB sinkless changefeed support that cuts out the Kafka middleman. DuckDB v2.0's async I/O shows a 20x speedup on CSV over S3 and 3x on Parquet. PostgreSQL 19's SQL/PGQ graph queries are real and useful for data lineage, but variable-length path support is missing and the syntax will put most people off.","url":"https://daily.dev/posts/dolt-squashes-history-10x-faster-debezium-3-7-adds-cockroachdb-sinkless-cdc-zkmhohqzx","datePublished":"2026-08-01T04:19:28.588Z","dateModified":"2026-08-01T04:19:50.206Z","author":{"@type":"Organization","name":"Databases Digest","logo":"https://media.daily.dev/image/upload/s--ir1hdV8b--/f_auto,q_auto/v1773839400/logos/databases_digest?_a=BAMAMiiu0","url":"https://daily.dev/sources/databases_digest"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/databases_digest","name":"Databases Digest"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Databases Digest","item":"https://daily.dev/sources/databases_digest"},{"@type":"ListItem","position":3,"name":"Dolt squashes history 10x faster, Debezium 3.7 adds CockroachDB sinkless CDC"}]} ```