gmp-bq-query

name	gmp-bq-query
version	1.0.0
description	BigQuery: Run pre-built query templates against GMP export datasets.
metadata	{"openclaw":{"category":"analytics","requires":{"bins":["gmp"]},"cliHelp":"gmp bq query --help"}}

bq query

PREREQUISITE: Read ../gmp-shared/SKILL.md for auth and ../gmp-bq/SKILL.md for BQ overview.

Run a pre-built query template against a BigQuery dataset containing GMP export data.

Usage

gmp bq query -t <template-name> --project <gcp-project> --dataset <dataset> [flags]

Flags

Flag	Required	Default	Description
-t, --template <name>	Yes	--	Template name (use gmp bq templates to list)
--project <id>	Yes	--	GCP project ID
--dataset <name>	Yes	--	BigQuery dataset name
-r, --date-range <range>	No	30d	Date range: 30d, 7d, 2024-01-01..2024-03-31
-f, --format <fmt>	No	json	Output format: json, table, csv

Template Parameters

All templates accept the same four parameters, populated automatically:

Parameter	Source
{{project}}	--project flag
{{dataset}}	--dataset flag
{{start_date}}	Computed from --date-range
{{end_date}}	Computed from --date-range

Examples

# GA4: Flatten events from last 30 days
gmp bq query -t ga4-events-flat --project my-gcp-project --dataset analytics_123456789

# GA4: Funnel analysis for last 90 days
gmp bq query -t ga4-funnel --project my-gcp-project --dataset analytics_123456789 -r 90d -f table

# GA4: Attribution report for Q1 2024
gmp bq query -t ga4-attribution --project my-gcp-project --dataset analytics_123456789 -r 2024-01-01..2024-03-31

# Ads: Find wasted spend in last 30 days
gmp bq query -t ads-wasted-spend --project my-gcp-project --dataset my_ads_data -r 30d -f table

# GSC: Top pages for last 90 days
gmp bq query -t gsc-top-pages --project my-gcp-project --dataset searchconsole -r 90d -f csv > pages.csv

# GSC: Detect keyword cannibalization
gmp bq query -t gsc-cannibalization --project my-gcp-project --dataset searchconsole -r 90d -f table

Output

All templates return tabular data compatible with formatOutput (json, table, csv). Each template produces different columns — run with -f json to see the full field list.

Tips

• Use gmp bq templates --service ga4 to see only GA4 templates.
• Use gmp bq explore --project X --dataset Y to discover available tables before running queries.
• GA4 dataset names follow the pattern analytics_<PROPERTY_ID>.
• Pipe JSON output to jq for filtering: gmp bq query ... | jq '[.[] | select(.clicks > 100)]'
• Export to CSV for spreadsheet analysis: gmp bq query ... -f csv > report.csv

See Also

• gmp-bq -- All BQ commands
• gmp-shared -- Global flags and auth