use big-endian representation for uuids

prestodb · Oct 16, 2024 · 347c794 · 347c794
1 parent 9fa4aee
commit 347c794
Show file tree

Hide file tree

Showing 2 changed files with 90 additions and 8 deletions.
diff --git a/presto-common/src/main/java/com/facebook/presto/common/type/UuidType.java b/presto-common/src/main/java/com/facebook/presto/common/type/UuidType.java
@@ -28,6 +28,7 @@
 import static com.facebook.presto.common.type.TypeSignature.parseTypeSignature;
 import static io.airlift.slice.SizeOf.SIZE_OF_LONG;
 import static io.airlift.slice.Slices.wrappedLongArray;
+import static java.lang.Long.reverseBytes;
 import static java.lang.String.format;
 
 public class UuidType
@@ -161,8 +162,8 @@ public final Slice getSlice(Block block, int position)
     public static Slice javaUuidToPrestoUuid(UUID uuid)
     {
         return wrappedLongArray(
-                uuid.getMostSignificantBits(),
-                uuid.getLeastSignificantBits());
+                reverseBytes(uuid.getMostSignificantBits()),
+                reverseBytes(uuid.getLeastSignificantBits()));
     }
 
     public static UUID prestoUuidToJavaUuid(Slice uuid)
@@ -171,7 +172,7 @@ public static UUID prestoUuidToJavaUuid(Slice uuid)
             throw new IllegalStateException(format("Expected value to be exactly %d bytes but was %d", INT128_BYTES, uuid.length()));
         }
         return new UUID(
-                uuid.getLong(0),
-                uuid.getLong(SIZE_OF_LONG));
+                reverseBytes(uuid.getLong(0)),
+                reverseBytes(uuid.getLong(SIZE_OF_LONG)));
     }
 }
diff --git a/presto-main/src/test/java/com/facebook/presto/type/TestUuidOperators.java b/presto-main/src/test/java/com/facebook/presto/type/TestUuidOperators.java
@@ -105,11 +105,92 @@ public void testNotEquals()
     @Test
     public void testOrderOperators()
     {
-        assertFunction("CAST('12151fd2-7586-11e9-8f9e-2a86e4085a58' AS UUID) < CAST('12151fd2-7586-11e9-8f9e-2a86e4085a59' AS UUID)", BOOLEAN, true);
-        assertFunction("CAST('12151fd2-7586-11e9-8f9e-2a86e4085a59' AS UUID) < CAST('12151fd2-7586-11e9-8f9e-2a86e4085a58' AS UUID)", BOOLEAN, false);
+        // LT
+        // equality
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) < CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        // check bytes in LSB
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) < CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) < CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-ff00-000000000001' AS UUID) < CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) < CAST('00000000-0000-0000-ff00-000000000000' AS UUID)", BOOLEAN, true);
 
-        assertFunction("UUID '12151fd2-7586-11e9-8f9e-2a86e4085a52' BETWEEN UUID '12151fd2-7586-11e9-8f9e-2a86e4085a50' AND UUID '12151fd2-7586-11e9-8f9e-2a86e4085a59'", BOOLEAN, true);
-        assertFunction("UUID '12151fd2-7586-11e9-8f9e-2a86e4085a52' BETWEEN UUID '12151fd2-7586-11e9-8f9e-2a86e4085a54' AND UUID '12151fd2-7586-11e9-8f9e-2a86e4085a59'", BOOLEAN, false);
+        // check bytes in MSB
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) < CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) < CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('ff000000-0000-0001-0000-000000000000' AS UUID) < CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) < CAST('ff000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+
+        // check MSB side takes precedent over LSB side
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) < CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) < CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('ff000000-0000-0000-0000-000000000001' AS UUID) < CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) < CAST('ff000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+
+        // ========================
+        // LTE
+        // equality
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) <= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        // check bytes in LSB
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) <= CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) <= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-ff00-000000000001' AS UUID) <= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) <= CAST('00000000-0000-0000-ff00-000000000000' AS UUID)", BOOLEAN, true);
+
+        // check bytes in MSB
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) <= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) <= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('ff000000-0000-0001-0000-000000000000' AS UUID) <= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) <= CAST('ff000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+
+        // check MSB side takes precedent over LSB side
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) <= CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) <= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('ff000000-0000-0000-0000-000000000001' AS UUID) <= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) <= CAST('ff000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+
+        // ========================
+        // GT
+        // equality
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) > CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+        // check bytes in LSB
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) > CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) > CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-ff00-000000000001' AS UUID) > CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) > CAST('00000000-0000-0000-ff00-000000000000' AS UUID)", BOOLEAN, false);
+
+        // check bytes in MSB
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) > CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) > CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('ff000000-0000-0001-0000-000000000000' AS UUID) > CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) > CAST('ff000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+
+        // check MSB side takes precedent over LSB side
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) > CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) > CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('ff000000-0000-0000-0000-000000000001' AS UUID) > CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) > CAST('ff000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
+
+        // ========================
+        // GTE
+        // equality
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) >= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        // check bytes in LSB
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) >= CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) >= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-ff00-000000000001' AS UUID) >= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) >= CAST('00000000-0000-0000-ff00-000000000000' AS UUID)", BOOLEAN, false);
+
+        // check bytes in MSB
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) >= CAST('00000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000000' AS UUID) >= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('ff000000-0000-0001-0000-000000000000' AS UUID) >= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) >= CAST('ff000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+
+        // check MSB side takes precedent over LSB side
+        assertFunction("CAST('00000000-0000-0001-0000-000000000000' AS UUID) >= CAST('00000000-0000-0000-0000-000000000001' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) >= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, false);
+        assertFunction("CAST('ff000000-0000-0000-0000-000000000001' AS UUID) >= CAST('00000000-0000-0001-0000-000000000000' AS UUID)", BOOLEAN, true);
+        assertFunction("CAST('00000000-0000-0000-0000-000000000001' AS UUID) >= CAST('ff000000-0000-0000-0000-000000000000' AS UUID)", BOOLEAN, false);
     }
 
     @Test