Skip to main content

CQL (Aster TSDB)

Time series data, such as tag values and events, is stored here. What this storage engine excels at is reading a single tag's data over a specific time range in milliseconds.

ItemValue
Port9042
Keyspacepp
Representative Tablespp.tm_tag_point · tm_tag_point_archive · tm_tag_point_snapshot · pp.ts_data_points · pp.tm_asset_data
EngineAster TSDB fork for industrial time series → Time Series Database

Connecting​

docker exec -it plantpulse-datalake pd node cql

🔴 CQL is not SQL​

Aster TSDB is a storage engine that finds data by partition key and scans in clustering order. Follow just this one rule — provide the partition key and time range in WHERE.

-- ✅ 이렇게
SELECT * FROM pp.tm_tag_point
 WHERE tag_id = 'TAG-001'
   AND timestamp >= '2026-07-01' AND timestamp < '2026-07-02';
-- ❌ 이렇게 하지 마십시오
SELECT * FROM pp.tm_tag_point WHERE value > 100 ALLOW FILTERING;
The ALLOW FILTERING is a warning, not a permission

The fact that Aster TSDB requires this means it must scan all nodes. Running this on a production cluster will slow down all other queries at that moment. Send such queries to Spark SQL.

Examples​

Recent values​

SELECT timestamp, value FROM pp.tm_tag_point
 WHERE tag_id = 'TAG-001'
 ORDER BY timestamp DESC
 LIMIT 20;

Time range slicing​

SELECT timestamp, value, quality FROM pp.tm_tag_point
 WHERE tag_id = 'TAG-001'
   AND timestamp >= '2026-07-01 08:00' AND timestamp < '2026-07-01 09:00';

Hourly average — computed on the server​

SELECT time_bucket(1h, timestamp) AS bucket, avg(value)
FROM pp.tm_tag_point
WHERE tag_id = 'TAG-001'
  AND timestamp >= '2026-07-01' AND timestamp < '2026-07-02'
GROUP BY tag_id, time_bucket(1h, timestamp);

21 time series functions, including time_bucket, are built in → Time Series Functions.

Log body search​

SELECT ts, msg FROM logs
 WHERE device = 'pump-01'
   AND ts >= '2026-07-31' AND ts < '2026-08-01'
   AND msg LIKE '%타임아웃%';

This works without ALLOW FILTERING → Full-Text Search.

Viewing the schema​

DESCRIBE KEYSPACES;
DESCRIBE TABLE pp.tm_tag_point;

-- 이 테이블에 걸린 컴팩션·계층화 설정
SELECT compaction, extensions FROM system_schema.tables
 WHERE keyspace_name = 'pp' AND table_name = 'tm_tag_point';

Things to know​

  • Sorting is only possible in the clustering key direction. You cannot sort by any arbitrary column using ORDER BY.
  • No joins. If you need to combine multiple tables, that's what Spark SQL is for.
  • Per-table TTL, compression strategy, and size are shown in pd retention.
  • Even if old data is compressed, queries are identical — SELECT handles the merging automatically → Tiered Storage.

See also​