Skip to content

The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:

Drop the nonexistent schema from Metabot's Sample Database examples (#82958)

Andrei Simionescue6dcb849bea7
diff --git a/resources/metabot/prompts/shared/prompt_snippets/mbql-shape.selmer b/resources/metabot/prompts/shared/prompt_snippets/mbql-shape.selmer
index ad8bc3f7c44..af0ea0f5c02 100644
--- a/resources/metabot/prompts/shared/prompt_snippets/mbql-shape.selmer
+++ b/resources/metabot/prompts/shared/prompt_snippets/mbql-shape.selmer
@@ -17,9 +17,9 @@ Questions of the form *"which / who / what has the most / highest / slowest / la
```json
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
- "aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]]],
- "breakout": [["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]]],
+ "source-table": ["Sample Database", null, "ORDERS"],
+ "aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]]],
+ "breakout": [["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]]],
"order-by": [["desc", {}, ["aggregation", {}, 0]]],
"limit": 1}
```
diff --git a/resources/metabot/prompts/tools/construct_notebook_query.md b/resources/metabot/prompts/tools/construct_notebook_query.md
index 05e6b12984b..0237812a1c4 100644
--- a/resources/metabot/prompts/tools/construct_notebook_query.md
+++ b/resources/metabot/prompts/tools/construct_notebook_query.md
@@ -13,10 +13,10 @@ Return:
```json
{"lib/type": "mbql/query",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]}]}
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]}]}
```
Every clause is `["op", {}, ...args]` with a mandatory `{}` options map at position 1; every field reference uses a 4-segment portable FK in the last slot. These are the two most-violated rules.
@@ -48,7 +48,7 @@ There is no top-level `database:` field in the LLM contract — the database is
["field", {}, ["<db-name>", "<schema-or-null>", "<table-name>", "<field-name>"]]
```
-The third slot is the **portable field FK** — a 4+ element string array. Schemaless databases (MongoDB, etc.) use `null` in the schema slot: `["Mongo", null, "orders", "created_at"]`. JSON-unfolded fields append extra segments: `["DB", "SCH", "TBL", "PARENT", "CHILD"]`.
+The third slot is the **portable field FK** — a 4+ element string array. Schemaless databases (the Sample Database, MongoDB, etc.) use `null` in the schema slot: `["Mongo", null, "orders", "created_at"]`. JSON-unfolded fields append extra segments: `["DB", "SCH", "TBL", "PARENT", "CHILD"]`.
Inside later stages, refer to a column produced by the previous stage by **string name** instead of a portable FK: `["field", {}, "count"]`, `["field", {}, "PRODUCT_ID"]`.
@@ -68,14 +68,14 @@ Filter (comparison + boolean combination):
```json
"filters": [["and", {},
- [">", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]], 100],
- ["=", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "STATUS"]], "paid"]]]
+ [">", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]], 100],
+ ["<=", {}, ["field", {}, ["Sample Database", null, "ORDERS", "QUANTITY"]], 5]]]
```
Aggregation (on a field, plus `count`):
```json
-"aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]],
+"aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]],
["count", {}]]
```
@@ -83,13 +83,13 @@ Breakout with temporal bucket:
```json
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]
```
Order by — direction wraps a ref; works on field refs or aggregation refs:
```json
-"order-by": [["desc", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]],
+"order-by": [["desc", {}, ["field", {}, ["Sample Database", null, "ORDERS", "CREATED_AT"]]],
["desc", {}, ["aggregation", {}, 0]]]
```
@@ -110,8 +110,8 @@ Define custom columns inside a stage using `expressions` and reference by name w
```json
"expressions": {
"Subtotal": ["+", {},
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]],
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TAX"]]]
+ ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]],
+ ["field", {}, ["Sample Database", null, "ORDERS", "TAX"]]]
},
"aggregation": [["sum", {}, ["expression", {}, "Subtotal"]]]
```
@@ -125,14 +125,14 @@ The sequential form `[["expression", {}, "Name", expr], ...]` is also accepted a
"alias": "Products",
"strategy": "left-join",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "PRODUCTS"]}],
+ "source-table": ["Sample Database", null, "PRODUCTS"]}],
"conditions": [["=", {},
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]],
+ ["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]],
["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "ID"]]]]
+ ["Sample Database", null, "PRODUCTS", "ID"]]]]
}],
"breakout": [["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "CATEGORY"]]]
+ ["Sample Database", null, "PRODUCTS", "CATEGORY"]]]
```
- `alias` is a free-choice string; use it consistently in conditions, breakout, aggregation, order-by.
@@ -145,9 +145,9 @@ The sequential form `[["expression", {}, "Name", expr], ...]` is also accepted a
Reference a field on a related table directly — when the source has exactly one FK to that target, the tool auto-fills `source-field` and performs the implicit join:
```json
-"source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+"source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
-"breakout": [["field", {}, ["Sample Database", "PUBLIC", "PRODUCTS", "CATEGORY"]]]
+"breakout": [["field", {}, ["Sample Database", null, "PRODUCTS", "CATEGORY"]]]
```
- **Multiple FKs** to the target → `:ambiguous-fk` lists them. Retry with explicit `{"source-field": ["DB", "SCH", "SRC", "FK_COL"]}` on the field.
@@ -167,9 +167,9 @@ Post-aggregation filter — count orders per product, keep only those with > 10:
```json
"stages": [
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
- "breakout": [["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]]]},
+ "breakout": [["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]]]},
{"lib/type": "mbql.stage/mbql",
"filters": [[">", {}, ["field", {}, "count"], 10]]}
]
@@ -180,10 +180,10 @@ Re-aggregate — average daily total by month:
```json
"stages": [
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
- "aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]]],
+ "source-table": ["Sample Database", null, "ORDERS"],
+ "aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]]],
"breakout": [["field", {"temporal-unit": "day"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]},
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]},
{"lib/type": "mbql.stage/mbql",
"aggregation": [["avg", {}, ["field", {}, "sum"]]],
"breakout": [["field", {"temporal-unit": "month"}, ["field", {}, "CREATED_AT"]]]}
diff --git a/resources/metabot/skills/construct-notebook-query-advanced.md b/resources/metabot/skills/construct-notebook-query-advanced.md
index 1270d3d884d..416bde0b7a4 100644
--- a/resources/metabot/skills/construct-notebook-query-advanced.md
+++ b/resources/metabot/skills/construct-notebook-query-advanced.md
@@ -24,8 +24,8 @@ Define custom columns inside a stage using `expressions` and reference by name w
```json
"expressions": {
"Subtotal": ["+", {},
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]],
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TAX"]]]
+ ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]],
+ ["field", {}, ["Sample Database", null, "ORDERS", "TAX"]]]
},
"aggregation": [["sum", {}, ["expression", {}, "Subtotal"]]]
```
@@ -39,14 +39,14 @@ A sequential form is also accepted, but the name must live in the defining claus
"alias": "Products",
"strategy": "left-join",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "PRODUCTS"]}],
+ "source-table": ["Sample Database", null, "PRODUCTS"]}],
"conditions": [["=", {},
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]],
+ ["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]],
["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "ID"]]]]
+ ["Sample Database", null, "PRODUCTS", "ID"]]]]
}],
"breakout": [["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "CATEGORY"]]]
+ ["Sample Database", null, "PRODUCTS", "CATEGORY"]]]
```
- `alias` is a free-choice string; use it consistently in conditions, breakout, aggregation, order-by.
@@ -59,9 +59,9 @@ A sequential form is also accepted, but the name must live in the defining claus
Reference a field on a related table directly — when the source has exactly one FK to that target, the tool auto-fills `source-field` and performs the implicit join:
```json
-"source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+"source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
-"breakout": [["field", {}, ["Sample Database", "PUBLIC", "PRODUCTS", "CATEGORY"]]]
+"breakout": [["field", {}, ["Sample Database", null, "PRODUCTS", "CATEGORY"]]]
```
- **Multiple FKs** to the target → `:ambiguous-fk` lists them. Retry with explicit `{"source-field": ["DB", "SCH", "SRC", "FK_COL"]}` on the field.
@@ -81,9 +81,9 @@ Post-aggregation filter — count orders per product, keep only those with > 10:
```json
"stages": [
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
- "breakout": [["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]]]},
+ "breakout": [["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]]]},
{"lib/type": "mbql.stage/mbql",
"filters": [[">", {}, ["field", {}, "count"], 10]]}
]
@@ -94,10 +94,10 @@ Re-aggregate — average daily total by month:
```json
"stages": [
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
- "aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]]],
+ "source-table": ["Sample Database", null, "ORDERS"],
+ "aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]]],
"breakout": [["field", {"temporal-unit": "day"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]},
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]},
{"lib/type": "mbql.stage/mbql",
"aggregation": [["avg", {}, ["field", {}, "sum"]]],
"breakout": [["field", {"temporal-unit": "month"}, ["field", {}, "CREATED_AT"]]]}
@@ -117,13 +117,13 @@ A rate like **open rate**, **bounce rate**, **conversion rate**, or **% of total
```json
{"lib/type": "mbql/query",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["/", {},
["count-where", {},
- ["<", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]], 50]],
+ ["<", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]], 50]],
["count", {}]]],
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]}]}
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]}]}
```
**Two stages** — use this when you also want the numerator and denominator as their own columns, or when the ratio needs further post-aggregation work. Compute them as separate aggregations in **stage 1**, then divide in **stage 2**.
@@ -134,12 +134,12 @@ Worked example — share of small orders per month (`count of orders under $50
{"lib/type": "mbql/query",
"stages": [
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [
- ["count-where", {}, ["<", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]], 50]],
+ ["count-where", {}, ["<", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]], 50]],
["count", {}]],
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]},
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]},
{"lib/type": "mbql.stage/mbql",
"expressions": {
"Small Order Rate": ["/", {}, ["field", {}, "count_where"], ["field", {}, "count"]]},
@@ -178,9 +178,9 @@ A metric is a pre-defined aggregation attached to a base table. To use one:
```json
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"filters": [[">", {}, ["field", {},
- ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]], 0]],
+ ["Sample Database", null, "ORDERS", "TOTAL"]], 0]],
"aggregation": [["metric", {}, "aB3cD4eF5gH6iJ7kL8mN9"]]}
```
@@ -192,18 +192,18 @@ To group a metric by a field that lives on a **different** table, add an explici
```json
{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["metric", {}, "aB3cD4eF5gH6iJ7kL8mN9"]],
"joins": [{"alias": "Products",
"strategy": "left-join",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "PRODUCTS"]}],
+ "source-table": ["Sample Database", null, "PRODUCTS"]}],
"conditions": [["=", {},
- ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "PRODUCT_ID"]],
+ ["field", {}, ["Sample Database", null, "ORDERS", "PRODUCT_ID"]],
["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "ID"]]]]}],
+ ["Sample Database", null, "PRODUCTS", "ID"]]]]}],
"breakout": [["field", {"join-alias": "Products"},
- ["Sample Database", "PUBLIC", "PRODUCTS", "CATEGORY"]]]}
+ ["Sample Database", null, "PRODUCTS", "CATEGORY"]]]}
```
When the breakout dimension lives on the metric's **own** base table, no join is needed — just break out on a portable FK as in the example above.
diff --git a/resources/metabot/skills/construct-notebook-query-core.md b/resources/metabot/skills/construct-notebook-query-core.md
index 7d92e70ae9e..f10c0d38a5b 100644
--- a/resources/metabot/skills/construct-notebook-query-core.md
+++ b/resources/metabot/skills/construct-notebook-query-core.md
@@ -26,10 +26,10 @@ Return:
```json
{"lib/type": "mbql/query",
"stages": [{"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
+ "source-table": ["Sample Database", null, "ORDERS"],
"aggregation": [["count", {}]],
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]}]}
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]}]}
```
Every clause is `["op", {}, ...args]` with a mandatory `{}` options map at position 1; every field reference uses a 4-segment portable FK in the last slot. These are the two most-violated rules.
@@ -61,7 +61,7 @@ There is no top-level `database:` field in the LLM contract — the database is
["field", {}, ["<db-name>", "<schema-or-null>", "<table-name>", "<field-name>"]]
```
-The third slot is the **portable field FK** — a 4+ element string array. Schemaless databases (MongoDB, etc.) use `null` in the schema slot: `["Mongo", null, "orders", "created_at"]`. JSON-unfolded fields append extra segments: `["DB", "SCH", "TBL", "PARENT", "CHILD"]`.
+The third slot is the **portable field FK** — a 4+ element string array. Schemaless databases (the Sample Database, MongoDB, etc.) use `null` in the schema slot: `["Mongo", null, "orders", "created_at"]`. JSON-unfolded fields append extra segments: `["DB", "SCH", "TBL", "PARENT", "CHILD"]`.
Inside later stages, refer to a column produced by the previous stage by **string name** instead of a portable FK: `["field", {}, "count"]`, `["field", {}, "PRODUCT_ID"]`.
@@ -81,14 +81,14 @@ Filter (comparison + boolean combination):
```json
"filters": [["and", {},
- [">", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]], 100],
- ["=", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "STATUS"]], "paid"]]]
+ [">", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]], 100],
+ ["<=", {}, ["field", {}, ["Sample Database", null, "ORDERS", "QUANTITY"]], 5]]]
```
Aggregation (on a field, plus `count`):
```json
-"aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]],
+"aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]],
["count", {}]]
```
@@ -96,13 +96,13 @@ Breakout with temporal bucket:
```json
"breakout": [["field", {"temporal-unit": "month"},
- ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]]
+ ["Sample Database", null, "ORDERS", "CREATED_AT"]]]
```
Order by — direction wraps a ref; works on field refs or aggregation refs:
```json
-"order-by": [["desc", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "CREATED_AT"]]],
+"order-by": [["desc", {}, ["field", {}, ["Sample Database", null, "ORDERS", "CREATED_AT"]]],
["desc", {}, ["aggregation", {}, 0]]]
```
diff --git a/resources/metabot/skills/read-resource.md b/resources/metabot/skills/read-resource.md
index 8009f34a2f5..15d9f386842 100644
--- a/resources/metabot/skills/read-resource.md
+++ b/resources/metabot/skills/read-resource.md
@@ -54,7 +54,8 @@ You can request multiple resources in one call by providing a list of URIs (max
- `metabase://database/{id}/schemas/{schemaName}/tables` — tables in a specific schema
**Examples:**
-- Want to see warehouse layout before writing SQL? → `metabase://database/1/schemas` then `metabase://database/1/schemas/PUBLIC/tables`
+- Want to see warehouse layout before writing SQL? → `metabase://database/2/schemas` then `metabase://database/2/schemas/public/tables`
+- Database without schemas, like the Sample Database (its `/schemas` list is empty)? → `metabase://database/1/tables`
- Want curated models in a specific warehouse? → `metabase://database/1/models`
**Best Practices:**
diff --git a/src/metabase/agent_api/api.clj b/src/metabase/agent_api/api.clj
index 28039648505..40ed93817cf 100644
--- a/src/metabase/agent_api/api.clj
+++ b/src/metabase/agent_api/api.clj
@@ -287,10 +287,10 @@
"```\n"
"{\"query\": {\"lib/type\": \"mbql/query\",\n"
" \"stages\": [{\"lib/type\": \"mbql.stage/mbql\",\n"
- " \"source-table\": [\"Sample Database\", \"PUBLIC\", \"ORDERS\"],\n"
+ " \"source-table\": [\"Sample Database\", null, \"ORDERS\"],\n"
" \"aggregation\": [[\"count\", {}]],\n"
" \"breakout\": [[\"field\", {\"temporal-unit\": \"month\"},\n"
- " [\"Sample Database\", \"PUBLIC\", \"ORDERS\", \"CREATED_AT\"]]]}]}}\n"
+ " [\"Sample Database\", null, \"ORDERS\", \"CREATED_AT\"]]]}]}}\n"
"```\n"
"\n"
"Common pitfalls: (1) forgetting the `{}` options map; (2) writing numeric "
diff --git a/src/metabase/agent_api/reference.md b/src/metabase/agent_api/reference.md
index 92400e8f79f..a5db24abbc4 100644
--- a/src/metabase/agent_api/reference.md
+++ b/src/metabase/agent_api/reference.md
@@ -206,7 +206,7 @@ Response:
"display_name": "Orders",
"description": "All customer orders",
"database_id": 1,
- "database_schema": "PUBLIC",
+ "database_schema": null,
"verified": false
},
{
@@ -256,9 +256,9 @@ Minimal example:
"stages": [
{
"lib/type": "mbql.stage/mbql",
- "source-table": ["Sample Database", "PUBLIC", "ORDERS"],
- "aggregation": [["sum", {}, ["field", {}, ["Sample Database", "PUBLIC", "ORDERS", "TOTAL"]]]],
- "breakout": [["field", {}, ["Sample Database", "PUBLIC", "PEOPLE", "STATE"]]],
+ "source-table": ["Sample Database", null, "ORDERS"],
+ "aggregation": [["sum", {}, ["field", {}, ["Sample Database", null, "ORDERS", "TOTAL"]]]],
+ "breakout": [["field", {}, ["Sample Database", null, "PEOPLE", "STATE"]]],
"limit": 100
}
]
diff --git a/src/metabase/metabot/tools/construct.clj b/src/metabase/metabot/tools/construct.clj
index 8c399615f9a..2c80b90cf49 100644
--- a/src/metabase/metabot/tools/construct.clj
+++ b/src/metabase/metabot/tools/construct.clj
@@ -54,9 +54,9 @@
"(possibly empty) options map at position 1, and every field reference is `[\"field\", {}, "
"[\"<db>\", \"<schema>\", \"<table>\", \"<field>\"]]`. Minimal example — count of orders by "
"month: `{\"lib/type\": \"mbql/query\", \"stages\": [{\"lib/type\": \"mbql.stage/mbql\", "
- "\"source-table\": [\"Sample Database\", \"PUBLIC\", \"ORDERS\"], \"aggregation\": "
+ "\"source-table\": [\"Sample Database\", null, \"ORDERS\"], \"aggregation\": "
"[[\"count\", {}]], \"breakout\": [[\"field\", {\"temporal-unit\": \"month\"}, "
- "[\"Sample Database\", \"PUBLIC\", \"ORDERS\", \"CREATED_AT\"]]]}]}`. Load the "
+ "[\"Sample Database\", null, \"ORDERS\", \"CREATED_AT\"]]]}]}`. Load the "
"`construct-notebook-query-*` skills for the full operator catalog, joins, expressions, "
"and multi-stage rules.")
:required ["lib/type" "stages"]
@@ -74,7 +74,7 @@
"source-table" {:type "array"
:minItems 3
:maxItems 3
- :items {:type "string"}
+ :items {}
:description "Portable FK `[<db-name>, <schema-or-null>, <table-name>]`."}
"source-card" {:type "string"
:description "entity_id of a saved question/model used as the source."}
diff --git a/test/metabase/metabot/agent/prompts_test.clj b/test/metabase/metabot/agent/prompts_test.clj
index 81ff091669b..f86a89eb4b8 100644
--- a/test/metabase/metabot/agent/prompts_test.clj
+++ b/test/metabase/metabot/agent/prompts_test.clj
@@ -3,11 +3,14 @@
[clojure.java.io :as io]
[clojure.string :as str]
[clojure.test :refer :all]
+ [metabase.agent-api.api]
[metabase.api-scope.core :as api-scope]
+ [metabase.api.macros :as api.macros]
[metabase.metabot.agent.profiles :as profiles]
[metabase.metabot.agent.prompts :as prompts]
[metabase.metabot.scope :as scope]
- [metabase.metabot.skills :as skills]))
+ [metabase.metabot.skills :as skills]
+ [metabase.metabot.tools.construct :as construct]))
(deftest ^:parallel load-system-prompt-template-test
(testing "loads internal.selmer template"
@@ -34,6 +37,19 @@
(is (str/includes? prompt "Sample Database"))
(is (not (re-find #"\[Sample\s*," prompt))))))
+(deftest ^:parallel sample-database-examples-use-a-null-schema-test
+ (doseq [[source text] {"construct_notebook_query.md" (slurp (io/resource "metabot/prompts/tools/construct_notebook_query.md"))
+ "construct-notebook-query-core.md" (slurp (io/resource "metabot/skills/construct-notebook-query-core.md"))
+ "construct-notebook-query-advanced.md" (slurp (io/resource "metabot/skills/construct-notebook-query-advanced.md"))
+ "mbql-shape.selmer" (slurp (io/resource "metabot/prompts/shared/prompt_snippets/mbql-shape.selmer"))
+ "construct_notebook_query tool" (:description @#'construct/construct-notebook-query-json-schema)
+ "Agent API construct_query tool" (-> (api.macros/find-route 'metabase.agent-api.api :post "/v2/construct-query")
+ (get-in [:form :metadata :tool :description]))
+ "Agent API reference.md" (slurp (io/resource "metabase/agent_api/reference.md"))}]
+ (testing source
+ (is (= #{"null"}
+ (set (map second (re-seq #"\"Sample Database\",\s*([^,\s]+)" text))))))))
+
(deftest ^:parallel render-system-prompt-test
(testing "renders template with variables"
(let [template "Hello {{name}}, today is {{day}}"