@gooddata/code-cli 0.50.0-alpha.5 → 0.50.0-alpha.6

package/dist/fixtures/rules/computedAttributes.mdc

@@ -0,0 +1,51 @@
+---
+description: Detailed information about GoodData computed attributes definition.
+alwaysApply: false
+---
+# Fields
+
+On top of generic fields like `id`, `type` (always "computed_attribute"), `title`, `description` and
+`tags`, computed attributes support:
+- `maql`: (string, required) a MAQL statement that assigns a group name to every row.
+- `locale`: (string) a locale whose collation order the values are sorted by.
+
+The `id` must be unique among computed attributes, and must not collide with the id of an attribute
+or a label in the same workspace.
+
+# What a computed attribute is
+
+A computed attribute turns numbers into named groups - for example "Small", "Medium", "Large" - and
+then behaves like any other attribute: it can be used to group a report, and it lives next to metrics
+rather than in the data model.
+
+# MAQL
+
+Write the break points as a CASE expression over a metric. String values use double quotes, and the
+first matching branch wins, so order the break points from the most specific one:
+
+```
+SELECT CASE WHEN {metric/won_activities} > 100 THEN "High" WHEN {metric/won_activities} > 50 THEN "Mid" ELSE "Low" END
+```
+
+Add a `BY` clause to pin the grain the metric is evaluated at. Without it the metric is evaluated in
+whatever context the report slices by, so the group a row falls into would change with the report:
+
+```
+SELECT CASE WHEN {metric/won_activities} > 100 THEN "High" ELSE "Low" END BY {label/sales_rep.name}
+```
+
+A computed attribute can be referenced from a metric, or from another computed attribute, under the
+same `computed_attribute` type it is declared with:
+
+```
+SELECT COUNT({computed_attribute/rep_performance})
+```
+
+MAQL syntax is different from SQL. Do not try generating MAQL expressions based on SQL knowledge.
+Instead, point user to GoodData documentation:
+https://www.gooddata.com/docs/cloud/create-metrics/maql/
+
+# Limits
+
+Only break points the user writes out are supported. Groups by fixed count (histograms) and groups by
+a fixed step are not.

package/dist/fixtures/rules/gooddata.mdc

@@ -37,5 +37,6 @@ A typical folder structure for GoodData analytics project looks like this:
 - `{{ANALYTICS}}/datasets` - a folder with all datasets and date instances
 - `{{ANALYTICS}}/metrics` - a folder with all metrics
 - `{{ANALYTICS}}/parameters` - a folder with all parameters
+- `{{ANALYTICS}}/computedAttributes` - a folder with all computed attributes
 - `{{ANALYTICS}}/visualisations` - a folder with all visualizations
 - `{{ANALYTICS}}/dashboards` - a folder with all dashboards