Add tag-based query filtering for IoTDB table model

[JackieTien97]

Summary

• Add ENABLE_QUERY_TAG_FILTER, QUERY_TAG_INDEX, and QUERY_TAG_VALUE_NUM configuration options with documented defaults and validation.
• Enumerate the selected TAG distinct values from benchmark-generated devices per target table, honoring REAL_INSERT_RATE; QUERY_TAG_VALUE_NUM=0 selects all actual values in that table, while positive values sample without replacement.
• Replace the device_id predicate with a TAG IN (...) predicate for Q1–Q12 in the IoTDB 2.0 table model, while keeping device and verification queries exact.
• Preserve legacy behavior by default and add workload, configuration, TAG-generation, SQL-construction, escaping, and end-to-end tests.

Why

The benchmark previously filtered query workloads only by device_id, so it could not exercise TAG-based filtering in the IoTDB table model. This change makes TAG filtering configurable and guarantees that generated IN values belong to the benchmark-configured value set for the selected query table.

Configuration

• ENABLE_QUERY_TAG_FILTER=false: enables TAG filtering for IoTDB 2.0 table-model queries.
• QUERY_TAG_INDEX=0: zero-based TAG column index used by the filter.
• QUERY_TAG_VALUE_NUM=0: number of distinct values in the IN list; zero means all actual distinct values for the selected TAG in the target table.

Invalid combinations fail fast, including unsupported database/dialect modes and requested counts larger than the available per-table cardinality.

Validation

• mvn -B spotless:check
• mvn -B test -pl iotdb-2.0 -am
  • Core: 167 tests passed
  • IoTDB 2.0: 11 tests passed