Optimize Trino SQL queries using TD-specific time functions for partition pruning.
npx skills add https://github.com/treasure-data/td-skills --skill trino
Please help me install this Agent Skill. Skill: trino Source: https://github.com/treasure-data/td-skills/tree/main/sql-skills/trino Command: npx skills add https://github.com/treasure-data/td-skills --skill trino
This Skill provides expert assistance for writing, analyzing, and optimizing Trino SQL queries for Treasure Data, helping users overcome performance bottlenecks, memory issues, and complex query logic. It ensures efficient data retrieval and processing within the TD environment, saving time and compute resources.
TD_INTERVAL, TD_TIME_RANGE), selecting specific columns, and leveraging approximate functions for large datasets.TD_INTERVAL, TD_TIME_STRING, and TD_SESSIONIZE for efficient time-series analysis.TD_INTERVAL and APPROX_DISTINCT, ensuring the query runs quickly and efficiently without hitting memory limits.SELECT TD_TIME_STRING(time, 'd!', 'JST') as date, APPROX_DISTINCT(user_id) as unique_users FROM database_name.events WHERE TD_INTERVAL(time, '-1w', 'JST') GROUP BY 1
High-intent search queries and answers about installing and using this skill.
Query optimization in Trino relies on partition pruning using time-based filtering. Use TD_INTERVAL to restrict queries to specific time ranges, select only required columns, and leverage APPROX_DISTINCT for large cardinality sets. This reduces data scans and memory consumption, accelerating execution.
TD_INTERVAL and TD_TIME_RANGE are Treasure Data SQL functions that filter data by time periods efficiently. TD_INTERVAL accepts relative ranges like '-1w' for the past week; TD_TIME_RANGE handles explicit date bounds. Both enable partition pruning, reducing query cost and runtime significantly.
Combine TD_TIME_STRING to format timestamps and APPROX_DISTINCT to count unique users, grouped by date. Apply TD_INTERVAL in the WHERE clause to filter the time range. This pattern efficiently aggregates user activity without full table scans.
Yes, TD_INTERVAL supports relative time syntax like '-1w', '-1d', or '-1mo' to filter data relative to query execution time. Specify your timezone to ensure consistent interpretation. This eliminates manual date calculation and supports dynamic time-window analysis.
Avoid using td_scheduled_time() as the first argument to td_interval; this prevents reliable partition pruning. Ensure timestamp columns have consistent timezones across your data. Use td_scheduled_time() as a reference date elsewhere in queries to enable pattern compatibility and optimization.
Use TD_TIME_STRING to convert Unix timestamps to readable formats with timezone support. Apply it in the SELECT clause after filtering with TD_INTERVAL, so partition pruning occurs before formatting. This balances readability with efficient data retrieval.
phanijappsCollect, normalize, and validate OHLCV market data from yfinance.
delta-and-betaExport Braze data from dashboard CSVs and Export API responses into analyzable datasets.
PremModhaOfficialRoute GROUP BY and time-series aggregation queries to DuckDB via pg_duckdb in PostgreSQL 18.
treasure-dataQuery Treasure Data via Presto or Hive and load pandas DataFrames.