[draft] Preserve underscores in custom schema name prefixes

.nextchanges/bundles/name-prefix-underscores.md

@@ -0,0 +1 @@
+* Fix underscores in custom `presets.name_prefix` for schemas, registered models, and serving endpoints; existing deployments using such prefixes may recreate these resources, and schema recreation can delete data, so inspect `bundle plan` before deploying. ([#7287](https://github.com/databricks/cli/pull/7287))

acceptance/bundle/presets/name_prefix_existing_schema/databricks.yml

@@ -0,0 +1,14 @@
+bundle:
+  name: name-prefix-existing-schema
+
+resources:
+  schemas:
+    qc:
+      name: qc
+      catalog_name: main
+
+targets:
+  existing:
+    default: true
+    presets:
+      name_prefix: my_user

acceptance/bundle/presets/name_prefix_existing_schema/out.test.toml

@@ -0,0 +1,2 @@
+Cloud = false
+EnvMatrix.DMS = ["", "true"]

acceptance/bundle/presets/name_prefix_existing_schema/output.txt

@@ -0,0 +1,20 @@
+
+>>> [CLI] bundle deploy
+Uploading bundle files to /Workspace/Users/[USERNAME]/.bundle/name-prefix-existing-schema/existing/files...
+Created schemas.qc
+Files: 1 uploaded, 0 deleted
+Resources: 1 created, 0 changed, 0 deleted, 0 unchanged
+
+>>> [CLI] bundle summary -o json
+"main.my_userqc"
+
+>>> [CLI] bundle plan
+Plan: 0 to add, 0 to change, 0 to delete, 1 unchanged
+
+=== A corrected prefix plans to recreate the existing schema
+>>> update_file.py databricks.yml name_prefix: my_user name_prefix: my_user_
+
+>>> [CLI] bundle plan
+recreate schemas.qc
+
+Plan: 1 to add, 0 to change, 1 to delete, 0 unchanged

acceptance/bundle/presets/name_prefix_existing_schema/script

@@ -0,0 +1,8 @@
+# Older CLIs resolve my_user_ to my_userqc, the name created here.
+trace $CLI bundle deploy
+trace $CLI bundle summary -o json | jq .resources.schemas.qc.id
+trace $CLI bundle plan
+
+title "A corrected prefix plans to recreate the existing schema"
+trace update_file.py databricks.yml "name_prefix: my_user" "name_prefix: my_user_"
+trace $CLI bundle plan

acceptance/bundle/presets/name_prefix_underscores/databricks.yml

@@ -0,0 +1,25 @@
+bundle:
+  name: name_prefix_underscores
+
+resources:
+  schemas:
+    qc:
+      name: qc
+      catalog_name: my_catalog
+  registered_models:
+    model:
+      name: model
+  model_serving_endpoints:
+    endpoint:
+      name: endpoint
+
+targets:
+  trailing:
+    default: true
+    presets:
+      name_prefix: ${workspace.current_user.short_name}_
+  repeated:
+    presets:
+      name_prefix: team__${workspace.current_user.short_name}__
+  generated:
+    mode: development

acceptance/bundle/presets/name_prefix_underscores/out.test.toml

@@ -0,0 +1,2 @@
+Cloud = false
+EnvMatrix.DMS = ["", "true"]

acceptance/bundle/presets/name_prefix_underscores/output.txt

@@ -0,0 +1,21 @@
+
+>>> [CLI] bundle validate -o json -t trailing
+{
+  "schema": "[USERNAME]_qc",
+  "model": "[USERNAME]_model",
+  "endpoint": "[USERNAME]_endpoint"
+}
+
+>>> [CLI] bundle validate -o json -t repeated
+{
+  "schema": "team__[USERNAME]__qc",
+  "model": "team__[USERNAME]__model",
+  "endpoint": "team__[USERNAME]__endpoint"
+}
+
+>>> [CLI] bundle validate -o json -t generated
+{
+  "schema": "dev_[USERNAME]_qc",
+  "model": "dev_[USERNAME]_model",
+  "endpoint": "dev_[USERNAME]_endpoint"
+}

acceptance/bundle/presets/name_prefix_underscores/script

@@ -0,0 +1,3 @@
+trace $CLI bundle validate -o json -t trailing | jq '{schema: .resources.schemas.qc.name, model: .resources.registered_models.model.name, endpoint: .resources.model_serving_endpoints.endpoint.name}'
+trace $CLI bundle validate -o json -t repeated | jq '{schema: .resources.schemas.qc.name, model: .resources.registered_models.model.name, endpoint: .resources.model_serving_endpoints.endpoint.name}'
+trace $CLI bundle validate -o json -t generated | jq '{schema: .resources.schemas.qc.name, model: .resources.registered_models.model.name, endpoint: .resources.model_serving_endpoints.endpoint.name}'

bundle/config/mutator/resourcemutator/apply_presets.go

@@ -372,18 +372,20 @@ func toTagArray(tags map[string]string) []Tag {
 	return tagArray
 }
 
-// normalizePrefix prefixes strings like '[dev lennart] ' to 'dev_lennart_'.
-// We leave unicode letters and numbers but remove all "special characters."
+// normalizePrefix normalizes special characters while preserving explicit underscores.
 func normalizePrefix(prefix string) string {
 	prefix = strings.ReplaceAll(prefix, "[", "")
 	prefix = strings.Trim(prefix, " ")
 
-	// If the prefix ends with a ']', we add an underscore to the end.
-	// This makes sure that we get names like "dev_user_endpoint" instead of "dev_userendpoint"
-	suffix := ""
-	if strings.HasSuffix(prefix, "]") {
-		suffix = "_"
-	}
+	bracketed := strings.HasSuffix(prefix, "]")
 
-	return textutil.NormalizeString(prefix) + suffix
+	parts := strings.Split(prefix, "_")
+	for i := range parts {
+		parts[i] = textutil.NormalizeString(parts[i])
+	}
+	normalized := strings.Join(parts, "_")
+	if bracketed && !strings.HasSuffix(normalized, "_") {
+		normalized += "_"
+	}
+	return normalized
 }

bundle/config/mutator/resourcemutator/apply_presets_test.go

@@ -7,6 +7,7 @@ import (
 	"github.com/databricks/cli/bundle/config"
 	"github.com/databricks/cli/bundle/config/resources"
 	"github.com/databricks/databricks-sdk-go/service/catalog"
+	"github.com/databricks/databricks-sdk-go/service/serving"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -38,3 +39,46 @@ func TestApplyPresetsSkipsNilQualityMonitors(t *testing.T) {
 	require.NoError(t, diags.Error())
 	assert.Nil(t, b.Config.Resources.QualityMonitors["monitor"].Schedule)
 }
+
+func TestApplyPresetsNormalizedPrefixPreservesUnderscores(t *testing.T) {
+	tests := []struct {
+		name       string
+		prefix     string
+		wantPrefix string
+	}{
+		{"trailing underscore", "my_user_", "my_user_"},
+		{"repeated underscores", "my__user__", "my__user__"},
+		{"bracketed underscores", "[qa__team] ", "qa__team_"},
+		{"bracketed trailing underscore", "[qa_] ", "qa_"},
+		{"bracketed repeated trailing underscores", "[qa__] ", "qa__"},
+		{"bracketed prefix", "[prod] ", "prod_"},
+		{"other separators", "a-b", "a_b"},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			b := &bundle.Bundle{
+				Config: config.Root{
+					Presets: config.Presets{NamePrefix: test.prefix},
+					Resources: config.Resources{
+						Schemas: map[string]*resources.Schema{
+							"qc": {CreateSchema: catalog.CreateSchema{CatalogName: "my_catalog", Name: "qc"}},
+						},
+						RegisteredModels: map[string]*resources.RegisteredModel{
+							"model": {CreateRegisteredModelRequest: catalog.CreateRegisteredModelRequest{Name: "model"}},
+						},
+						ModelServingEndpoints: map[string]*resources.ModelServingEndpoint{
+							"endpoint": {CreateServingEndpoint: serving.CreateServingEndpoint{Name: "endpoint"}},
+						},
+					},
+				},
+			}
+
+			diags := bundle.Apply(t.Context(), b, ApplyPresets())
+			require.NoError(t, diags.Error())
+			assert.Equal(t, test.wantPrefix+"qc", b.Config.Resources.Schemas["qc"].Name)
+			assert.Equal(t, test.wantPrefix+"model", b.Config.Resources.RegisteredModels["model"].Name)
+			assert.Equal(t, test.wantPrefix+"endpoint", b.Config.Resources.ModelServingEndpoints["endpoint"].Name)
+		})
+	}
+}