Indexed columns first in WHERE

uniq() for distinct counts

quantile() for percentiles

No SELECT *

No FINAL keyword

No excessive JOINs

sumMerge/countMerge/uniqMerge for MV queries

system.query_log monitoring

system.parts for table size

ClickHouse time functions

countIf() for conditional aggregation

Uses 'clickhouse' package

Port 8123

basicAuth configuration

Environment variables for credentials

Batch insert for bulk loading

No individual inserts in loop

Streaming insert for continuous data

Batch size consideration

toPromise() or async/await

No SELECT * in any queries

MergeTree engine used

index_granularity setting

PARTITION BY toYYYYMM

ORDER BY indexed columns first

ReplacingMergeTree for deduplication

PRIMARY KEY on ReplacingMergeTree

LowCardinality for repeated strings

Enum for categorical data

Smallest appropriate integer type

No SELECT * in queries

Time-based partitioning only

AggregatingMergeTree or Materialized View