DAT-18792: added cluster by support for snapshot and diff table related change types

src/main/java/liquibase/ext/databricks/diff/output/changelog/MissingTableChangeGeneratorDatabricks.java

@@ -14,7 +14,6 @@
 
 public class MissingTableChangeGeneratorDatabricks extends MissingTableChangeGenerator {
 
-
     @Override
     public int getPriority(Class<? extends DatabaseObject> objectType, Database database) {
         if (database instanceof DatabricksDatabase && Table.class.isAssignableFrom(objectType)) {
@@ -35,12 +34,14 @@ public Change[] fixMissing(DatabaseObject missingObject, DiffOutputControl contr
         ExtendedTableProperties extendedTableProperties = new ExtendedTableProperties(
                 null,
                 missingObject.getAttribute("tblProperties", String.class));
+        String clusterColumns = missingObject.getAttribute("clusteringColumns", "");
 
-        changes[0] = getCreateTableChangeDatabricks(extendedTableProperties, changes);
+        changes[0] = getCreateTableChangeDatabricks(extendedTableProperties, changes, clusterColumns);
         return changes;
     }
 
-    private CreateTableChangeDatabricks getCreateTableChangeDatabricks(ExtendedTableProperties extendedTableProperties, Change[] changes) {
+    private CreateTableChangeDatabricks getCreateTableChangeDatabricks(ExtendedTableProperties extendedTableProperties,
+                                                                       Change[] changes, String clusterColumns) {
         CreateTableChange temp = (CreateTableChange) changes[0];
         CreateTableChangeDatabricks createTableChangeDatabricks = new CreateTableChangeDatabricks();
         createTableChangeDatabricks.setColumns(temp.getColumns());
@@ -52,6 +53,9 @@ private CreateTableChangeDatabricks getCreateTableChangeDatabricks(ExtendedTable
         createTableChangeDatabricks.setRemarks(temp.getRemarks());
         createTableChangeDatabricks.setIfNotExists(temp.getIfNotExists());
         createTableChangeDatabricks.setRowDependencies(temp.getRowDependencies());
+        if (!clusterColumns.isEmpty()) {
+            createTableChangeDatabricks.setClusterColumns(clusterColumns);
+        }
 
         createTableChangeDatabricks.setExtendedTableProperties(extendedTableProperties);
         return createTableChangeDatabricks;
@@ -61,4 +65,4 @@ private CreateTableChangeDatabricks getCreateTableChangeDatabricks(ExtendedTable
     protected CreateTableChange createCreateTableChange() {
         return new CreateTableChangeDatabricks();
     }
-}
+}

src/main/java/liquibase/ext/databricks/snapshot/jvm/TableSnapshotGeneratorDatabricks.java

@@ -13,12 +13,14 @@
 
 import java.util.List;
 import java.util.Map;
+import java.util.regex.Pattern;
+import java.util.stream.Collectors;
 
 public class TableSnapshotGeneratorDatabricks extends TableSnapshotGenerator {
 
     private static final String LOCATION = "Location";
-    private static final String TABLE_PROPERTIES = "Table Properties";
     private static final String TBL_PROPERTIES = "tblProperties";
+    private static final String CLUSTER_COLUMNS = "clusteringColumns";
     private static final String DETAILED_TABLE_INFORMATION_NODE = "# Detailed Table Information";
 
     @Override
@@ -49,11 +51,13 @@ protected DatabaseObject snapshotObject(DatabaseObject example, DatabaseSnapshot
                 if (detailedInformationNode && tableProperty.get("COL_NAME").equals(LOCATION)) {
                     table.setAttribute(LOCATION, tableProperty.get("DATA_TYPE"));
                 }
-                if (detailedInformationNode && tableProperty.get("COL_NAME").equals(TABLE_PROPERTIES)) {
-                    String tblProperties = (String) tableProperty.get("DATA_TYPE");
-                    table.setAttribute(TBL_PROPERTIES, tblProperties.substring(1, tblProperties.length() - 1));// remove starting and ending square brackets
-                }
             }
+            Map<String, String> tblProperties = getTblPropertiesMap(database, example.getName());
+            if (tblProperties.containsKey(CLUSTER_COLUMNS)) {
+                // removing clusterColumns, as clusterColumns tblProperty is not allowed in create/alter table statements
+                table.setAttribute(CLUSTER_COLUMNS, sanitizeClusterColumns(tblProperties.remove(CLUSTER_COLUMNS)));
+            }
+            table.setAttribute(TBL_PROPERTIES, getTblPropertiesString(tblProperties));
         }
         return table;
     }
@@ -63,4 +67,26 @@ protected DatabaseObject snapshotObject(DatabaseObject example, DatabaseSnapshot
     // Location = get column 'storage_path'
     // cleanup this after approach of getting all properties is settled
 
+    private Map<String, String> getTblPropertiesMap(Database database, String table) throws DatabaseException {
+        String query = String.format("SHOW TBLPROPERTIES %s.%s.%s;", database.getDefaultCatalogName(), database.getDefaultSchemaName(), table);
+        List<Map<String, ?>> tablePropertiesResponse = Scope.getCurrentScope().getSingleton(ExecutorService.class)
+                .getExecutor("jdbc", database).queryForList(new RawParameterizedSqlStatement(query));
+        return tablePropertiesResponse.stream()
+                .collect(Collectors.toMap(mapElement -> (String) mapElement.get("KEY"), mapElement -> (String) mapElement.get("VALUE")));
+    }
+
+    private String getTblPropertiesString(Map<String, String> propertiesMap) {
+        StringBuilder csvString = new StringBuilder();
+        propertiesMap.entrySet()
+                .stream()
+                .sorted(Map.Entry.comparingByKey())
+                .forEach(entry -> csvString.append(entry.getKey()).append("=").append(entry.getValue()).append(","));
+        return csvString.toString().replaceAll(",$", "");
+    }
+
+    private String sanitizeClusterColumns(String clusterColumnProperty) {
+        Pattern pattern = Pattern.compile("[\\[\\]\\\"]");
+        return clusterColumnProperty.replaceAll(pattern.toString(), "");
+    }
+
 }

src/main/java/liquibase/ext/databricks/sqlgenerator/CreateTableGeneratorDatabricks.java

@@ -17,7 +17,6 @@
 
 public class CreateTableGeneratorDatabricks extends CreateTableGenerator {
 
-
     @Override
     public int getPriority() {
         return PRIORITY_DATABASE;

src/test/resources/liquibase/harness/change/changelogs/databricks/createClusteredTableNew.json

@@ -20,9 +20,15 @@
                     "name": "test_new",
                     "type": "int"
                   }
+                },
+                {
+                  "column": {
+                    "name": "test_present_new",
+                    "type": "int"
+                  }
                 }
               ],
-              "clusterColumns": "test_id,test_new"
+              "clusterColumns": "test_id,test_new,test_present_new"
             }
           }
         ],
@@ -48,6 +54,11 @@
                   "column": {
                     "name": "test_id"
                   }
+                },
+                {
+                  "column": {
+                    "name": "test_present_new`"
+                  }
                 }
               ]
             }

src/test/resources/liquibase/harness/change/changelogs/databricks/createClusteredTableNew.xml

@@ -13,7 +13,8 @@
         <createTable tableName="test_table_clustered_new" >
             <column name="test_id" type="int" />
             <column name="test_new" type="int"/>
-            <databricks:clusterColumns>test_id,test_new</databricks:clusterColumns>
+            <column name="test_present_new" type="int"/>
+            <databricks:clusterColumns>test_id,test_new,test_present_new</databricks:clusterColumns>
         </createTable>
         <rollback>
             <!-- The dropTable will drop a full table whether it has clustered columns or not. -->
@@ -23,6 +24,7 @@
     <changeSet id="2" author="your.name">
         <databricks:alterCluster tableName="test_table_clustered_new">
             <databricks:column name="test_id"/>
+            <databricks:column name="test_present_new"/>
         </databricks:alterCluster>
         <rollback/>
     </changeSet>

src/test/resources/liquibase/harness/change/changelogs/databricks/createClusteredTableNew.yaml

@@ -12,7 +12,10 @@ databaseChangeLog:
               - column:
                   name: test_new
                   type: int
-            clusterColumns: test_id, test_new
+              - column:
+                  name: test_present_new
+                  type: int
+            clusterColumns: test_id, test_new, test_present_new
       rollback:
         dropTable:
           tableName: test_table_clustered_new
@@ -25,6 +28,8 @@ databaseChangeLog:
             columns:
               - column:
                   name: test_id
+              - column:
+                  name: test_present_new
       rollback:
         empty
   - changeSet:

src/test/resources/liquibase/harness/change/expectedSnapshot/databricks/createClusteredTableNew.json

@@ -1,2 +1,26 @@
 {
+  "snapshot": {
+    "objects": {
+      "liquibase.structure.core.Table": [
+        {
+          "table": {
+            "name": "test_table_clustered_new",
+            "clusteringColumns": "test_id,test_present_new"
+          }
+        }
+      ],
+      "liquibase.structure.core.Column": [
+        {
+          "column": {
+            "name": "test_id"
+          }
+        },
+        {
+          "column": {
+            "name": "test_present_new"
+          }
+        }
+      ]
+    }
+  }
 }

src/test/resources/liquibase/harness/change/expectedSql/databricks/createClusteredTableNew.sql

@@ -1,3 +1,3 @@
-CREATE TABLE main.liquibase_harness_test_ds.test_table_clustered_new (test_id INT, test_new INT) USING delta TBLPROPERTIES('delta.feature.allowColumnDefaults' = 'supported', 'delta.columnMapping.mode' = 'name', 'delta.enableDeletionVectors' = true) CLUSTER BY (test_id, test_new)
-ALTER TABLE main.liquibase_harness_test_ds.test_table_clustered_new CLUSTER BY (test_id)
+CREATE TABLE main.liquibase_harness_test_ds.test_table_clustered_new (test_id INT, test_new INT, test_present_new INT) USING delta TBLPROPERTIES('delta.feature.allowColumnDefaults' = 'supported', 'delta.columnMapping.mode' = 'name', 'delta.enableDeletionVectors' = true) CLUSTER BY (test_id, test_new, test_present_new)
+ALTER TABLE main.liquibase_harness_test_ds.test_table_clustered_new CLUSTER BY (test_id,test_present_new)
 ALTER TABLE main.liquibase_harness_test_ds.test_table_clustered_new DROP COLUMN test_new