You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@phoenix.apache.org by GitBox <gi...@apache.org> on 2019/03/29 21:37:54 UTC

[GitHub] [phoenix] gjacoby126 commented on a change in pull request #473: PHOENIX-5138 - ViewIndexId sequences created after PHOENIX-5132 shoul…

gjacoby126 commented on a change in pull request #473: PHOENIX-5138 - ViewIndexId sequences created after PHOENIX-5132 shoul…
URL: https://github.com/apache/phoenix/pull/473#discussion_r270581181
 
 

 ##########
 File path: phoenix-core/src/it/java/org/apache/phoenix/end2end/UpgradeIT.java
 ##########
 @@ -588,4 +603,85 @@ public void testMoveParentChildLinks() throws Exception {
         return childLinkSet;
     }
 
+    @Test
+    public void testMergeViewIndexSequences() throws Exception {
+        testMergeViewIndexSequencesHelper(false);
+    }
+
+    @Test
+    public void testMergeViewIndexSequencesWithNamespaces() throws Exception {
+        testMergeViewIndexSequencesHelper(true);
+    }
+
+    private void testMergeViewIndexSequencesHelper(boolean isNamespaceMappingEnabled) throws Exception {
+        PhoenixConnection conn = getConnection(false, null, isNamespaceMappingEnabled).unwrap(PhoenixConnection.class);
+        ConnectionQueryServices cqs = conn.getQueryServices();
+        //First delete any sequences that may exist from previous tests
+        conn.createStatement().execute("DELETE FROM " + PhoenixDatabaseMetaData.SYSTEM_SEQUENCE);
+        conn.commit();
+        cqs.clearCache();
+        //Now make sure that running the merge logic doesn't cause a problem when there are no
+        //sequences
+        UpgradeUtil.mergeViewIndexIdSequences(cqs, conn);
+        PName tenantOne = PNameFactory.newName("TENANT_ONE");
+        PName tenantTwo = PNameFactory.newName("TENANT_TWO");
+        String tableName =
+            SchemaUtil.getPhysicalHBaseTableName("TEST",
+                "T_" + generateUniqueName(), isNamespaceMappingEnabled).getString();
+        PName viewIndexTable = PNameFactory.newName(MetaDataUtil.getViewIndexPhysicalName(tableName));
+        SequenceKey sequenceOne =
+            createViewIndexSequenceWithOldName(cqs, tenantOne, viewIndexTable, isNamespaceMappingEnabled);
+        SequenceKey sequenceTwo =
+            createViewIndexSequenceWithOldName(cqs, tenantTwo, viewIndexTable, isNamespaceMappingEnabled);
+        SequenceKey sequenceGlobal =
+            createViewIndexSequenceWithOldName(cqs, null, viewIndexTable, isNamespaceMappingEnabled);
+
+        List<SequenceAllocation> allocations = Lists.newArrayList();
+        long val1 = 10;
+        long val2 = 100;
+        long val3 = 1000;
+        allocations.add(new SequenceAllocation(sequenceOne, val1));
+        allocations.add(new SequenceAllocation(sequenceGlobal, val2));
+        allocations.add(new SequenceAllocation(sequenceTwo, val3));
+
+
+        long[] incrementedValues = new long[3];
+        SQLException[] exceptions = new SQLException[3];
+        cqs.incrementSequences(allocations, EnvironmentEdgeManager.currentTimeMillis(), incrementedValues,
+            exceptions);
+        for (SQLException e : exceptions) {
+            assertNull(e);
+        }
+
+        UpgradeUtil.mergeViewIndexIdSequences(cqs, conn);
+        //now check that there exists a sequence using the new naming convention, whose value is the
+        //max of all the previous sequences for this table.
+
+        List<SequenceAllocation> afterUpgradeAllocations = Lists.newArrayList();
+        SequenceKey sequenceUpgrade = MetaDataUtil.getViewIndexSequenceKey(null, viewIndexTable, 0, isNamespaceMappingEnabled);
+        afterUpgradeAllocations.add(new SequenceAllocation(sequenceUpgrade, 1));
+        long[] afterUpgradeValues = new long[1];
+        SQLException[] afterUpgradeExceptions = new SQLException[1];
+        cqs.incrementSequences(afterUpgradeAllocations, EnvironmentEdgeManager.currentTimeMillis(), afterUpgradeValues, afterUpgradeExceptions);
+
+        assertNull(afterUpgradeExceptions[0]);
+        if (isNamespaceMappingEnabled){
+            //since one sequence (the global one) will be reused as the "new" sequence,
 
 Review comment:
   For some historical reason, the sequence names were very different for namespace-mapped base tables, and in one case it coincides with the new naming convention. So in that case the "new" sequence is an old sequence that's been incremented an extra time, hence the extra 1

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services