The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:
TypeScript analyticsbibusiness-intelligencebusinessintelligenceclojuredashboarddatadata-analysisdata-visualizationdatabasemetabasemysqlpostgrespostgresqlreportingslacksql-editorvisualization
Stop translating agent-facing error prose (#85299)
Chris Vandoodewaardd94210f626cf
diff --git a/src/metabase/agent_lib/representations/repair.clj b/src/metabase/agent_lib/representations/repair.cljindex 2c6d61c6280..5df978cfd57 100644--- a/src/metabase/agent_lib/representations/repair.clj+++ b/src/metabase/agent_lib/representations/repair.clj@@ -44,7 +44,6 @@[metabase.models.serialization.resolve.mp :as resolve.mp][metabase.util :as u][metabase.util.date-2 :as u.date]- [metabase.util.i18n :refer [tru]][metabase.util.log :as log][metabase.util.malli.registry :as mr][metabase.util.match :as match]@@ -1851,8 +1850,8 @@(and inferred-type (not (contains? opts "effective-type")))(assoc "effective-type" inferred-type))]["aggregation" new-opts uuid])- (throw (ex-info (tru "Aggregation index {0} out of range; stage has {1} aggregation(s): {2}"- idx (count index->uuid) (agg-heads-summary aggs-for-error))+ (throw (ex-info (format "Aggregation index %s out of range; stage has %s aggregation(s): %s"+ idx (count index->uuid) (agg-heads-summary aggs-for-error)){:agent-error? true:error :aggregation-ref-out-of-range:index idx@@ -1884,7 +1883,7 @@;; Integer ref(s) present but stage has no aggregation vector - agent-error.(not (and (vector? aggs) (seq aggs)))- (throw (ex-info (tru "Found an integer-indexed aggregation reference but this stage has no `aggregation:` clause")+ (throw (ex-info "Found an integer-indexed aggregation reference but this stage has no `aggregation:` clause"{:agent-error? true:error :aggregation-ref-no-aggregations}))@@ -2101,7 +2100,7 @@(some? clause)))(map stage ["order-by" "limit" "fields"]))(throw (ex-info- (tru "Detected a post-aggregation filter alongside `order-by:`, `limit:`, or `fields:` in the same stage. MBQL 5 requires the filter in a separate stage, but those trailing clauses cannot be safely relocated — they may reference pre-aggregation columns that don''t survive the split. Please author this as two stages explicitly: stage 0 with `source-table`, `aggregation`, `breakout`, and any pre-aggregation `filters`; stage 1 with the post-aggregation `filters`, plus `order-by` / `limit` / `fields` that reference the aggregation''s output via cross-stage `[\"field\", (empty opts), \"<column-name>\"]` refs.")+ "Detected a post-aggregation filter alongside `order-by:`, `limit:`, or `fields:` in the same stage. MBQL 5 requires the filter in a separate stage, but those trailing clauses cannot be safely relocated — they may reference pre-aggregation columns that don't survive the split. Please author this as two stages explicitly: stage 0 with `source-table`, `aggregation`, `breakout`, and any pre-aggregation `filters`; stage 1 with the post-aggregation `filters`, plus `order-by` / `limit` / `fields` that reference the aggregation's output via cross-stage `[\"field\", (empty opts), \"<column-name>\"]` refs."{:agent-error? true:status-code 400:error-code :post-agg-filter-with-trailing-clauses@@ -2120,7 +2119,7 @@"filters" moved-filters}][s0 s1])(throw (ex-info- (tru "Detected a post-aggregation filter (a filter referencing an aggregation from the same stage). MBQL 5 requires this in a separate stage. Auto-fix is unsafe here — either an aggregation''s output column name cannot be derived (e.g. a `metric` reference or unknown head), or two aggregations in this stage would produce the same column name (e.g. two `sum` aggregations on different fields, which lib disambiguates as `sum` / `sum_2` based on field order). Please refactor to multi-stage manually: put the aggregations in stage 0, and the filter in stage 1 referencing the column via a cross-stage `[\"field\", (empty opts), \"<column-name>\"]` clause. Use an explicit aggregation with a `name` opts override (e.g. `[\"sum\", (opts with key \"name\" → \"sum_total\"), …]`) if you need stable cross-stage names.")+ "Detected a post-aggregation filter (a filter referencing an aggregation from the same stage). MBQL 5 requires this in a separate stage. Auto-fix is unsafe here — either an aggregation's output column name cannot be derived (e.g. a `metric` reference or unknown head), or two aggregations in this stage would produce the same column name (e.g. two `sum` aggregations on different fields, which lib disambiguates as `sum` / `sum_2` based on field order). Please refactor to multi-stage manually: put the aggregations in stage 0, and the filter in stage 1 referencing the column via a cross-stage `[\"field\", (empty opts), \"<column-name>\"]` clause. Use an explicit aggregation with a `name` opts override (e.g. `[\"sum\", (opts with key \"name\" → \"sum_total\"), …]`) if you need stable cross-stage names."{:agent-error? true:status-code 400:error-code :post-agg-filter-needs-multi-stage@@ -2605,10 +2604,10 @@;; so listing them in the diagnostic is fine. The FK column portables are NOT;; surfaced here — see the rationale on `:ambiguous-fk` above.(let [aliases (mapv :alias cands)]- (throw (ex-info (tru "Field {0} is reachable from {1} explicit joins via aliases {2}. Specify the `source-field-join-alias` option on the field clause to pick which join''s foreign key to use."- (display-portable fk)- (count cands)- (str/join ", " (map pr-str aliases)))+ (throw (ex-info (format "Field %s is reachable from %s explicit joins via aliases %s. Specify the `source-field-join-alias` option on the field clause to pick which join's foreign key to use."+ (display-portable fk)+ (count cands)+ (str/join ", " (map pr-str aliases))){:status-code 400:error :ambiguous-fk-via-join:agent-error? true@@ -3013,13 +3012,13 @@(defn- unresolved-join-alias-fields-error[problems](ex-info- (tru "Some join column references name a column their join doesn''t return: {0}. Reference a join column by its machine name (e.g. an aggregation is `min`, `count`, `sum` - never its display label like `Min of X`), and only columns the join returns: a grouped join returns just its breakouts and aggregations."- (str/join "; " (for [{:keys [stage join-alias columns available]} problems]- (tru "stage {0}, join {1} doesn''t return {2} (available: {3})"- stage- (quoting/quoted join-alias)- (str/join ", " (map quoting/quoted columns))- (str/join ", " (map quoting/quoted available))))))+ (format "Some join column references name a column their join doesn't return: %s. Reference a join column by its machine name (e.g. an aggregation is `min`, `count`, `sum` - never its display label like `Min of X`), and only columns the join returns: a grouped join returns just its breakouts and aggregations."+ (str/join "; " (for [{:keys [stage join-alias columns available]} problems]+ (format "stage %s, join %s doesn't return %s (available: %s)"+ stage+ (quoting/quoted join-alias)+ (str/join ", " (map quoting/quoted columns))+ (str/join ", " (map quoting/quoted available)))))){:agent-error? true:error :unresolved-join-alias-field:problems problems}))@@ -3120,9 +3119,9 @@(pos? idx) (mini-resolved-columns mp query idx content-store))](when-let [bad (first-unresolved-cross-stage-ref stage cols)](throw (ex-info- (tru "No column named {0} is available to reference in this stage. Reference a column by its machine name (e.g. an aggregation is `max`, `count`, `sum` - never its display label like `Max of X`), not a UI label. Available column names: {1}."- (quoting/quoted (nth bad 2))- (str/join ", " (map (comp quoting/quoted :name) cols)))+ (format "No column named %s is available to reference in this stage. Reference a column by its machine name (e.g. an aggregation is `max`, `count`, `sum` - never its display label like `Max of X`), not a UI label. Available column names: %s."+ (quoting/quoted (nth bad 2))+ (str/join ", " (map (comp quoting/quoted :name) cols))){:agent-error? true:error :unresolved-cross-stage-field:stage idx@@ -3169,8 +3168,8 @@(>= (count entry) 1)(= "field" (nth entry 0)))(throw (ex-info- (tru "Stage {0} aggregation entry {1} is a `field` clause, not an aggregation. The `aggregation:` block expects aggregation functions like `count`, `sum`, `avg`, `metric`, etc. - or wrap the field in one (e.g. `[sum, <opts>, <field>]`). If you wanted to group by this column, use `breakout:` instead."- stage-idx entry-idx)+ (format "Stage %s aggregation entry %s is a `field` clause, not an aggregation. The `aggregation:` block expects aggregation functions like `count`, `sum`, `avg`, `metric`, etc. - or wrap the field in one (e.g. `[sum, <opts>, <field>]`). If you wanted to group by this column, use `breakout:` instead."+ stage-idx entry-idx){:agent-error? true:error :aggregation-entry-not-aggregation:stage-index stage-idx@@ -3195,7 +3194,7 @@(match/match-one form[#{"case" "if"} {"default" _} & _](throw (ex-info- (tru "`case` (and `if`) uses its third positional argument as the fallback value, not a `default` key in the options map. Move the value out of the options map and append it as the third arg of the clause: `[case, <opts>, <branch-pairs>, <default>]`. Omit the third arg entirely if you have no fallback (the result will be null on miss).")+ "`case` (and `if`) uses its third positional argument as the fallback value, not a `default` key in the options map. Move the value out of the options map and append it as the third arg of the clause: `[case, <opts>, <branch-pairs>, <default>]`. Omit the third arg entirely if you have no fallback (the result will be null on miss)."{:agent-error? true:error :case-default-in-opts:clause &match}))))@@ -3226,8 +3225,8 @@(match/match-one form[(head :guard sexp-legacy-top-level-ops) & _](throw (ex-info- (tru "`{0}` is not a clause in repr; it was a top-level operation in the older sexp pipeline. {1}"- head (get sexp-legacy-top-level-ops head))+ (format "`%s` is not a clause in repr; it was a top-level operation in the older sexp pipeline. %s"+ head (get sexp-legacy-top-level-ops head)){:agent-error? true:error :sexp-legacy-op-as-clause:head head@@ -3247,7 +3246,7 @@(match/match-one form["expression" #'map? (name-slot :guard (or (not (string? name-slot)) (str/blank? name-slot)))](throw (ex-info- (tru "`[expression, <opts>, <name>]` reference requires a non-blank string identifier in the third slot, matching an entry in some stage''s `expressions:` block.")+ "`[expression, <opts>, <name>]` reference requires a non-blank string identifier in the third slot, matching an entry in some stage's `expressions:` block."{:agent-error? true:error :blank-expression-ref:clause &match}))))@@ -3270,7 +3269,7 @@(match/match-one form["field" #'map? #'integer?](throw (ex-info- (tru "`field` clause needs a portable FK in its third slot, not a numeric id. Use a vector `[<database>, <schema>, <table>, <column>]` (resolved against the metadata provider) or a string column-name (for cross-stage references).")+ "`field` clause needs a portable FK in its third slot, not a numeric id. Use a vector `[<database>, <schema>, <table>, <column>]` (resolved against the metadata provider) or a string column-name (for cross-stage references)."{:agent-error? true:error :numeric-field-id:clause &match})))))@@ -3342,8 +3341,8 @@(some? (get (nth node 1) "unit")))false))(throw (ex-info- (tru "Move the temporal unit in `{0}` off the literal. Write a range with `>=` the first day of the period and `<` the first day after the period, both as plain date strings (not `between`, which drops everything after midnight on its last day). To exclude `2025-01-31`, use an `or` of `< \"2025-01-31\"` and `>= \"2025-02-01\"` on the column. These comparisons drop rows with no date; add `is-null` on the column to the `or` if they should count. Join `conditions:` cannot hold an `or`, so a date exclusion cannot be written there; tell the user that it cannot be added to the join. You can instead put the unit on the column with `temporal-unit` in `filters:` and `expressions:`, then compare the column to the first day of the period as a plain date string. Do not put the unit on the column in `aggregation:` or join `conditions:`, where the query engine does not turn it into a range. Use the range for `year` on an `expression` reference, which the query engine reads as a number."- (pr-str node))+ (format "Move the temporal unit in `%s` off the literal. Write a range with `>=` the first day of the period and `<` the first day after the period, both as plain date strings (not `between`, which drops everything after midnight on its last day). To exclude `2025-01-31`, use an `or` of `< \"2025-01-31\"` and `>= \"2025-02-01\"` on the column. These comparisons drop rows with no date; add `is-null` on the column to the `or` if they should count. Join `conditions:` cannot hold an `or`, so a date exclusion cannot be written there; tell the user that it cannot be added to the join. You can instead put the unit on the column with `temporal-unit` in `filters:` and `expressions:`, then compare the column to the first day of the period as a plain date string. Do not put the unit on the column in `aggregation:` or join `conditions:`, where the query engine does not turn it into a range. Use the range for `year` on an `expression` reference, which the query engine reads as a number."+ (pr-str node)){:agent-error? true:error :unencodable-temporal-clause:clause node})))@@ -3547,8 +3546,8 @@:aggregation "`aggregation:`":filter "`filters:`")](throw (ex-info- (tru "The query builder''s expression editor rejects a {0} clause in stage {1}: {2}. The query would build but could not be visualized or saved - rewrite the offending clause."- where original-idx (str/replace message #"\.$" ""))+ (format "The query builder's expression editor rejects a %s clause in stage %s: %s. The query would build but could not be visualized or saved - rewrite the offending clause."+ where original-idx (str/replace message #"\.$" "")){:agent-error? true:error :expression-editor-rejection:status-code 400@@ -3605,10 +3604,10 @@type-name (name column-type)unit-name (name temporal-unit)](throw (ex-info (if (= :expression (nth clause 0))- (tru "{0} is a {1} custom column, so it cannot be bucketed by {2}. Compare it to a plain value, or bucket a date / datetime column instead."- column-name type-name unit-name)- (tru "{0} is a {1} column, so it cannot be bucketed by {2}. Compare it to a plain value, or bucket a date / datetime column instead."- column-name type-name unit-name))+ (format "%s is a %s custom column, so it cannot be bucketed by %s. Compare it to a plain value, or bucket a date / datetime column instead."+ column-name type-name unit-name)+ (format "%s is a %s column, so it cannot be bucketed by %s. Compare it to a plain value, or bucket a date / datetime column instead."+ column-name type-name unit-name)){:agent-error? true:error :temporal-unit-on-non-temporal-column:status-code 400diff --git a/src/metabase/metabot/tools/construct.clj b/src/metabase/metabot/tools/construct.cljindex a739a5e9404..a0ec40341f9 100644--- a/src/metabase/metabot/tools/construct.clj+++ b/src/metabase/metabot/tools/construct.clj@@ -35,7 +35,6 @@[metabase.models.serialization.resolve.mp :as resolve.mp][metabase.query-processor.core :as qp][metabase.util :as u]- [metabase.util.i18n :refer [tru]][metabase.util.json :as json][metabase.util.malli :as mu][metabase.util.malli.registry :as mr]@@ -193,7 +192,7 @@(let [raw (get-in parsed-query ["stages" 0 "source-table"])](when (string? raw)(when-let [[_ entity-type entity-id] (re-find metabase-uri-source-table-pattern raw)]- (throw (ex-info (tru "`source-table:` does not accept URIs like `{0}`." raw)+ (throw (ex-info (format "`source-table:` does not accept URIs like `%s`." raw){:agent-error? true:status-code 400:error :uri-in-source-table@@ -458,7 +457,7 @@(cond(pos-int? ref)(when-not (visible-table metadata-provider ref)- (throw (ex-info (tru "No table found with id {0}." (str ref))+ (throw (ex-info (format "No table found with id %s." (str ref)){:agent-error? true:status-code 400:error :unknown-table-id@@ -467,7 +466,7 @@(and (vector? ref) (= ::field-id (nth ref 0 nil)))(let [field-id (nth ref 1)](when-not (visible-field metadata-provider field-id)- (throw (ex-info (tru "No field found with id {0}." (str field-id))+ (throw (ex-info (format "No field found with id %s." (str field-id)){:agent-error? true:status-code 400:error :unknown-field-id@@ -477,8 +476,8 @@(let [[_ model id] ref](when-not (resolve-table-fk metadata-provider nil ref)(throw (ex-info (if (= model :model/Measure)- (tru "No measure found with id {0}." (str id))- (tru "No segment found with id {0}." (str id)))+ (format "No measure found with id %s." (str id))+ (format "No segment found with id %s." (str id))){:agent-error? true:status-code 400:error (if (= model :model/Measure)@@ -514,7 +513,7 @@(api/query-check :model/Table table-id)(catch clojure.lang.ExceptionInfo e(if (= 403 (:status-code (ex-data e)))- (throw (ex-info (tru "No table found with id {0}." (str table-id))+ (throw (ex-info (format "No table found with id %s." (str table-id)){:agent-error? true:status-code 400:error :unknown-table-id@@ -575,14 +574,14 @@(let [db-name (nth source-table-fk 0)ids (metabot.db/database-ids-by-name db-name)](case (count ids)- 0 (throw (ex-info (tru "Unknown database: `{0}`." db-name)+ 0 (throw (ex-info (format "Unknown database: `%s`." db-name){:agent-error? true:status-code 400:error :unknown-database:database db-name}))1 (first ids)- (throw (ex-info (tru "Multiple databases share the name `{0}` (ids: {1})."- db-name (pr-str (vec (sort ids))))+ (throw (ex-info (format "Multiple databases share the name `%s` (ids: %s)."+ db-name (pr-str (vec (sort ids)))){:agent-error? true:status-code 400:error :ambiguous-database-name@@ -595,12 +594,10 @@;; decides which provider to build), so an unchecked existence answer here would let the;; id argument enumerate table ids across databases the caller cannot otherwise see.(or (metabot.db/readable-active-table-database-id numeric-table)- ;; the id renders via str, not as a number — tru's MessageFormat would add locale- ;; digit grouping ("999,999,999").;; `:unknown-table-id`, not `:unknown-table`: a portable-FK miss (`:unknown-table`);; recovers by naming a portable FK, which is wrong advice for a numeric-id caller.- (throw (ex-info (tru "No table found with id {0}."- (str numeric-table))+ (throw (ex-info (format "No table found with id %s."+ (str numeric-table)){:agent-error? true:status-code 400:error :unknown-table-id@@ -609,7 +606,7 @@source-card-eid(if-let [card (tools.u/get-card-by-entity-id source-card-eid)](:database_id card)- (throw (ex-info (tru "No saved question or model found with entity_id {0}." (pr-str source-card-eid))+ (throw (ex-info (format "No saved question or model found with entity_id %s." (pr-str source-card-eid)){:agent-error? true:status-code 400:error :unknown-card@@ -631,8 +628,8 @@false(throw e)))))](when-not readable- (throw (ex-info (tru "No saved question or model found with id {0}."- (str card-id))+ (throw (ex-info (format "No saved question or model found with id %s."+ (str card-id)){:agent-error? true:status-code 400;; the numeric-specific key, matching `export-card-by-id` — a numeric@@ -642,7 +639,7 @@(:database_id card)):else- (throw (ex-info (tru "First stage must have either `source-table:` or `source-card:`; neither was found in `stages[0]`.")+ (throw (ex-info "First stage must have either `source-table:` or `source-card:`; neither was found in `stages[0]`."{:agent-error? true:status-code 400:error :missing-source-in-first-stage})))))@@ -850,17 +847,17 @@;; normalization stamps `:lib/uuid` on every clause; one missing means normalizing a clause;; inside it failed, often because it references something that doesn't exist(= :lib/uuid missing)- (tru "a reference inside it could not be resolved - often a field reference to a column, join alias, or expression that does not exist. Refer to a join''s or an earlier stage''s columns by their exact machine name (an aggregation is `min`, `count`, `sum`)")+ "a reference inside it could not be resolved - often a field reference to a column, join alias, or expression that does not exist. Refer to a join's or an earlier stage's columns by their exact machine name (an aggregation is `min`, `count`, `sum`)"(= :base-type missing)- (tru "a field reference by name is missing `base-type`, usually because the name matches no real column")+ "a field reference by name is missing `base-type`, usually because the name matches no real column"missing- (tru "missing `{0}`" (name missing))+ (format "missing `%s`" (name missing));; the message of `::lib.schema.expression/expression.definition`'s aggregation check(= "non-aggregation expression" (me/error-message error))- (tru "{0}: aggregation and window functions such as `offset` cannot be custom columns - move them to `aggregation:` or `order-by:`" (me/error-message error)))))+ (format "%s: aggregation and window functions such as `offset` cannot be custom columns - move them to `aggregation:` or `order-by:`" (me/error-message error)))))(defn- query-not-runnable-explanation"Explain why `pmbql-query` fails the query builder's `canRun` check, or return nil.@@ -879,9 +876,9 @@(take 3)seq(str/join "; "))- (tru "the query does not match the query builder''s schema"))+ "the query does not match the query builder's schema")(not-any? not-runnable-problem errors)- (str ". " (tru "This usually means a field reference is missing its type or names a column that does not exist, or an aggregation or window function such as `offset` is in `expressions:` instead of `aggregation:` or `order-by:`"))))))))+ (str ". This usually means a field reference is missing its type or names a column that does not exist, or an aggregation or window function such as `offset` is in `expressions:` instead of `aggregation:` or `order-by:`")))))))(defn- incompatible-metrics"Return the required sources of incompatible metrics in the first stage, sorted by metric ID.@@ -1050,8 +1047,8 @@_validated (repr/validate-query repaired)pmbql-query (repr.resolve/resolve-query mp repaired permission-aware-content-store)_runnable (when-let [why (query-not-runnable-explanation pmbql-query repaired)]- (throw (ex-info (tru "The constructed query is not runnable - it would fail the query builder''s validation, so it cannot be visualized or saved. Failing paths: {0}."- why)+ (throw (ex-info (format "The constructed query is not runnable - it would fail the query builder's validation, so it cannot be visualized or saved. Failing paths: %s."+ why){:agent-error? true:error :query-not-runnable:status-code 400})))diff --git a/src/metabase/metabot/tools/recovery_hints.clj b/src/metabase/metabot/tools/recovery_hints.cljindex 2e700023e61..ef1ef2fa5b8 100644--- a/src/metabase/metabot/tools/recovery_hints.clj+++ b/src/metabase/metabot/tools/recovery_hints.clj@@ -12,8 +12,7 @@which is the intended failure mode: a missing hint reads as terse, a wrong one sends theagent at a tool it does not have."(:require- [metabase.agent-lib.representations :as repr]- [metabase.util.i18n :refer [tru]]))+ [metabase.agent-lib.representations :as repr]))(set! *warn-on-reflection* true)@@ -29,13 +28,13 @@"`aggregation: [[metric, {}, \"<portable_entity_id>\"]]`.")("question" "model" "card")- (tru "To reference a saved question or model as a query source, put its `portable_entity_id` (the 21-char string from its search result or `read_resource`) into `source-card:` — not a URI.")+ "To reference a saved question or model as a query source, put its `portable_entity_id` (the 21-char string from its search result or `read_resource`) into `source-card:` — not a URI.""table"- (tru "Use the portable FK `[<db-name>, <schema>, <table-name>]` in `source-table:` — not a URI. `read_resource metabase://table/{0}` reports the exact names."- (str entity-id))+ (format "Use the portable FK `[<db-name>, <schema>, <table-name>]` in `source-table:` — not a URI. `read_resource metabase://table/%s` reports the exact names."+ (str entity-id))- (tru "`source-table:` accepts a portable FK `[<db-name>, <schema>, <table-name>]` or, via `source-card:`, a saved-card `portable_entity_id`.")))+ "`source-table:` accepts a portable FK `[<db-name>, <schema>, <table-name>]` or, via `source-card:`, a saved-card `portable_entity_id`."))(defn- card-fields-hint"Return guidance for listing `what` from the source card, or nil when `data` names no card."@@ -49,7 +48,7 @@[humanized](let [paths (repr/humanized->paths humanized)](when (seq paths)- (tru "Invalid at {0}. Fix the named paths." paths))))+ (format "Invalid at %s. Fix the named paths." paths))))(defn recovery-hint"The v1 recovery sentence for an agent error's `ex-data`, or nil when it has none."@@ -62,52 +61,52 @@(uri-hint entity-type entity-id):unknown-table- (tru "Call `read_resource` with `metabase://database/<numeric id>/tables` to list available tables and schemas, then retry with an exact portable FK from the response.")+ "Call `read_resource` with `metabase://database/<numeric id>/tables` to list available tables and schemas, then retry with an exact portable FK from the response.":unknown-table-id;; A numeric-id miss, not a portable-FK miss: pointing the agent at a portable FK (as;; `:unknown-table` does) would be wrong advice. v1 authors portable FKs, so the right move;; is to find the table by name and use its portable FK.- (tru "That numeric table id does not resolve. Call `read_resource` with `metabase://database/<numeric id>/tables` to find the table, then use its portable FK `[<db-name>, <schema>, <table-name>]` in `source-table:`.")+ "That numeric table id does not resolve. Call `read_resource` with `metabase://database/<numeric id>/tables` to find the table, then use its portable FK `[<db-name>, <schema>, <table-name>]` in `source-table:`.":ambiguous-table- (tru "Call `read_resource` with `metabase://database/<numeric id>/tables` to list available tables and retry with a more specific portable FK.")+ "Call `read_resource` with `metabase://database/<numeric id>/tables` to list available tables and retry with a more specific portable FK."(:unknown-field :unknown-field-id)- (tru "Call `read_resource` with `metabase://table/<numeric id>/fields` to list this table''s columns.")+ "Call `read_resource` with `metabase://table/<numeric id>/fields` to list this table's columns.":ambiguous-field- (tru "The field name matches more than one column in this table, so it cannot be referenced by name. Tell the user about the ambiguous column name.")+ "The field name matches more than one column in this table, so it cannot be referenced by name. Tell the user about the ambiguous column name.":ambiguous-fk(or (card-fields-hint data "foreign-key columns")- (tru "Call `read_resource` with `metabase://table/<numeric id>/fields` for the source table to list the available foreign-key columns."))+ "Call `read_resource` with `metabase://table/<numeric id>/fields` for the source table to list the available foreign-key columns.")(:column-not-returned :unexportable-fk)(card-fields-hint data "columns"):no-fk-path- (tru "If a metric relates to that table, read its dimensions resource `metabase://metric/<metric_id>/dimensions`, which lists the exact `joins:` clause to paste and the columns it unlocks.")+ "If a metric relates to that table, read its dimensions resource `metabase://metric/<metric_id>/dimensions`, which lists the exact `joins:` clause to paste and the columns it unlocks."(:unknown-card :unknown-card-id)- (tru "Do not invent or guess entity_ids: call `read_resource` with `metabase://question/<numeric id>` or `metabase://model/<numeric id>` first, then copy the exact `portable_entity_id` from the response into `source-card:`.")+ "Do not invent or guess entity_ids: call `read_resource` with `metabase://question/<numeric id>` or `metabase://model/<numeric id>` first, then copy the exact `portable_entity_id` from the response into `source-card:`."(:unknown-measure :unknown-measure-id)- (tru "Do not invent or guess entity_ids: read the table that owns the measure with `read_resource` (`metabase://table/<numeric id>`) and copy the exact `portable_entity_id` from its `<measure>` tag.")+ "Do not invent or guess entity_ids: read the table that owns the measure with `read_resource` (`metabase://table/<numeric id>`) and copy the exact `portable_entity_id` from its `<measure>` tag."(:unknown-segment :unknown-segment-id)- (tru "Do not invent or guess entity_ids: read the table that owns the segment with `read_resource` (`metabase://table/<numeric id>`) and copy the exact `portable_entity_id` from its `<segment>` tag.")+ "Do not invent or guess entity_ids: read the table that owns the segment with `read_resource` (`metabase://table/<numeric id>`) and copy the exact `portable_entity_id` from its `<segment>` tag."(:cross-database-card :cross-database-measure :cross-database-segment)- (tru "A query can only reference saved questions, models, metrics, measures and segments from its own database. Use one from that database, or change the source to a table in the referenced item''s database.")+ "A query can only reference saved questions, models, metrics, measures and segments from its own database. Use one from that database, or change the source to a table in the referenced item's database.":unknown-database- (tru "Use the exact database name as reported by search / `read_resource` (it appears as the first element of every portable FK, e.g. `source-table: [<db-name>, <schema>, <table>]`).")+ "Use the exact database name as reported by search / `read_resource` (it appears as the first element of every portable FK, e.g. `source-table: [<db-name>, <schema>, <table>]`).":ambiguous-database-name- (tru "Ask the user to rename one of the databases.")+ "Ask the user to rename one of the databases.":missing-source-in-first-stage- (tru "`source-table:` takes a portable FK `[<db-name>, <schema>, <table>]`; `source-card:` takes an entity_id string.")+ "`source-table:` takes a portable FK `[<db-name>, <schema>, <table>]`; `source-card:` takes an entity_id string."nil))