kater query compare

Compares two queries, or every query of two workspaces, and says where their rows differ.

Use it to check that a changed query still gives the same numbers. It runs the same comparison as kater compare, from a JSON request.

Examples

# Check that two queries give the same spend for each platform, with the request in compare.json
kater query compare --request-file compare.json
# Compare every query of ad_reporting with the same query in your brainstorm copy
echo '{"left": {"workspace": "connections/snowflake_prod_connection/workspaces/ad_reporting"}, "right": {"workspace": "connections/snowflake_prod_connection/users/you@example.com/brainstorm/promote_roas/workspaces/ad_reporting"}}' | kater query compare --request-file -

Arguments

Argument	Description
`<path>`	Optional. In place of `--request-file`: the workspace, the query file or the chart file the request names.

Options

Option	Description
`--request-file <file>`	The JSON request. `-` reads it from standard input. See The request.

It also takes the global options --json, --kater-root, --scope and --help.

The request

This request compares spend_by_platform with spend_conversion_value_by_platform, row by row for each platform and currency:

{"left": {"workspace": "connections/snowflake_prod_connection/workspaces/ad_reporting", "query": "spend_by_platform"}, "right": {"workspace": "connections/snowflake_prod_connection/workspaces/ad_reporting", "query": "spend_conversion_value_by_platform"}, "key": ["platform", "union__currency"]}

Key	What it holds
`left`, `right`	The two sides. A side is a saved query, `{"workspace": ..., "query": ...}`, or a scratch statement, `{"path": ..., "statement": ...}`, counted from 0. A `{"workspace": ...}` alone on both sides pairs the queries of two workspaces by name. Paths are from the content root.
`key`	The columns that pair a row of one side with a row of the other, by the left side’s names. Without it, Kater uses the row dimensions that both sides return. With none, it compares the sides as sets of rows.
`columns`	Left columns to compare with a right column of another name, as an object. Each key is a left name, and its value is the right name.
`parameters`	Parameter values by name, for each side that declares the parameter. Otherwise each side runs at its defaults, with its own filters on.
`precision`	The decimal places to compare numbers at, from 0 to 12, when that is finer than their format. Without it, numbers compare at their format’s places.

How it runs

Each side runs where you can already run it: a saved query as kater run runs it, and a scratch statement as the Sandbox runs it. Each pair runs as one warehouse statement that returns only the rows that differ, one pair at a time.