Skip to content

Set ORC columns and fix case-sensitivity issue with schema check #119

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
rzhang10 merged 1 commit into from
Aug 9, 2022
Merged

Set ORC columns and fix case-sensitivity issue with schema check #119

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
37 changes: 32 additions & 5 deletions ...src/main/java/org/apache/iceberg/hivelink/core/HiveMetadataPreservingTableOperations.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@
import org.apache.hadoop.hive.metastore.api.EnvironmentContext;
import org.apache.hadoop.hive.metastore.api.FieldSchema;
import org.apache.hadoop.hive.metastore.api.NoSuchObjectException;
import org.apache.hadoop.hive.metastore.api.SerDeInfo;
import org.apache.hadoop.hive.metastore.api.Table;
import org.apache.hadoop.hive.serde2.SerDeException;
import org.apache.hadoop.hive.serde2.avro.AvroObjectInspectorGenerator;
Expand All @@ -53,6 +54,7 @@
import org.apache.iceberg.relocated.com.google.common.base.Strings;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableSet;
import org.apache.iceberg.relocated.com.google.common.collect.Maps;
import org.apache.thrift.TException;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
Expand Down Expand Up @@ -82,6 +84,8 @@ public class HiveMetadataPreservingTableOperations extends HiveTableOperations {
.impl(HiveMetaStoreClient.class, "alter_table",
String.class, String.class, Table.class, EnvironmentContext.class)
.build();
public static final String ORC_COLUMNS = "columns";
public static final String ORC_COLUMNS_TYPES = "columns.types";

protected HiveMetadataPreservingTableOperations(Configuration conf, HiveClientPool metaClients, FileIO fileIO,
String catalogName, String database, String table) {
Expand Down Expand Up @@ -224,29 +228,30 @@ protected void doCommit(TableMetadata base, TableMetadata metadata) {
/**
* [LINKEDIN] Due to an issue that the table read in is sometimes corrupted and has incorrect columns, compare the
* table columns to the avro.schema.literal property (if it exists) and fix the table columns if there is a mismatch
* @return true if the schema was mismatched and fixed
*/
static void fixMismatchedSchema(Table table) {
static boolean fixMismatchedSchema(Table table) {
String avroSchemaLiteral = getAvroSchemaLiteral(table);
if (Strings.isNullOrEmpty(avroSchemaLiteral)) {
return;
return false;
}
Schema schema = new Schema.Parser().parse(avroSchemaLiteral);
List<FieldSchema> hiveCols;
try {
hiveCols = getColsFromAvroSchema(schema);
} catch (SerDeException e) {
LOG.error("Failed to get get columns from avro schema when checking schema", e);
return;
return false;
}

boolean schemaMismatched;
if (table.getSd().getCols().size() != hiveCols.size()) {
schemaMismatched = true;
} else {
Map<String, String> hiveFieldMap = hiveCols.stream().collect(
Collectors.toMap(FieldSchema::getName, FieldSchema::getType));
Collectors.toMap(field -> field.getName().toLowerCase(), field -> field.getType().toLowerCase()));
Map<String, String> tableFieldMap = table.getSd().getCols().stream().collect(
Collectors.toMap(FieldSchema::getName, FieldSchema::getType));
Collectors.toMap(field -> field.getName().toLowerCase(), field -> field.getType().toLowerCase()));
schemaMismatched = !hiveFieldMap.equals(tableFieldMap);
}

Expand All @@ -256,7 +261,13 @@ static void fixMismatchedSchema(Table table) {
table.getSd().getCols().stream().map(Object::toString).collect(Collectors.joining(", ")),
hiveCols.stream().map(Object::toString).collect(Collectors.joining(", ")));
table.getSd().setCols(hiveCols);
if (!Strings.isNullOrEmpty(table.getSd().getInputFormat()) && table.getSd().getInputFormat()
.contains("OrcInputFormat")) {
updateORCStorageDesc(hiveCols, table);
}
}

return schemaMismatched;
}

private static List<FieldSchema> getColsFromAvroSchema(Schema schema)
Expand All @@ -282,6 +293,22 @@ private static String getAvroSchemaLiteral(Table table) {
return schemaStr;
}

private static void updateORCStorageDesc(List<FieldSchema> hiveCols, Table table) {
String columnsString = hiveCols.stream().map(FieldSchema::getName).collect(Collectors.joining(","));
String typesString = hiveCols.stream().map(FieldSchema::getType).collect(Collectors.joining(","));

if (!table.getSd().isSetSerdeInfo()) {
table.getSd().setSerdeInfo(new SerDeInfo());
}
if (!table.getSd().getSerdeInfo().isSetParameters()) {
table.getSd().getSerdeInfo().setParameters(Maps.newHashMap());
}

Map<String, String> sdParams = table.getSd().getSerdeInfo().getParameters();
sdParams.put(ORC_COLUMNS, columnsString);
sdParams.put(ORC_COLUMNS_TYPES, typesString);
}

/**
* [LINKEDIN] a log-enhanced persistTable as a refactoring inspired by
* org.apache.iceberg.hive.HiveTableOperations#persistTable
Expand Down
15 changes: 14 additions & 1 deletion ...test/java/org/apache/iceberg/hivelink/core/TestHiveMetadataPreservingTableOperations.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,10 @@ public void testFixMismatchedSchema() {
"\"fields\":[{\"name\":\"name\",\"type\":\"string\"},{\"name\":\"id\",\"type\":\"int\"},{\"name\":\"nested\"," +
"\"type\":{\"name\":\"nested\",\"type\":\"record\",\"fields\":[{\"name\":\"field1\",\"type\":\"string\"}," +
"{\"name\":\"field2\",\"type\":\"string\"}]}}]}";
String testSchemaLiteralWithUppercase = "{\"name\":\"testSchema\",\"type\":\"record\",\"namespace\":\"com" +
".linkedin.test\", \"fields\":[{\"name\":\"Name\",\"type\":\"string\"},{\"name\":\"ID\",\"type\":\"int\"}" +
",{\"name\":\"Nested\", \"type\":{\"name\":\"Nested\",\"type\":\"record\",\"fields\":[{\"name\":\"Field1\"," +
"\"type\":\"string\"}, {\"name\":\"Field2\",\"type\":\"string\"}]}}]}";

long currentTimeMillis = System.currentTimeMillis();
StorageDescriptor storageDescriptor = new StorageDescriptor();
Expand All @@ -49,6 +53,7 @@ public void testFixMismatchedSchema() {
// Set cols with incorrect nested type
storageDescriptor.setCols(ImmutableList.of(field1, field2, new FieldSchema("nested", "struct<field1:int," +
"field2:string>", "")));
storageDescriptor.setInputFormat("org.apache.hadoop.hive.ql.io.orc.OrcInputFormat");
Map<String, String> parameters = ImmutableMap.of("avro.schema.literal", testSchemaLiteral);
Table tbl = new Table("tableName",
"dbName",
Expand All @@ -63,10 +68,18 @@ public void testFixMismatchedSchema() {
null,
TableType.EXTERNAL_TABLE.toString());

HiveMetadataPreservingTableOperations.fixMismatchedSchema(tbl);
Assert.assertTrue(HiveMetadataPreservingTableOperations.fixMismatchedSchema(tbl));
Assert.assertEquals(3, tbl.getSd().getColsSize());
Assert.assertEquals(field1, tbl.getSd().getCols().get(0));
Assert.assertEquals(field2, tbl.getSd().getCols().get(1));
Assert.assertEquals(field3, tbl.getSd().getCols().get(2));
Assert.assertTrue(storageDescriptor.getSerdeInfo().getParameters()
.containsKey(HiveMetadataPreservingTableOperations.ORC_COLUMNS));
Assert.assertTrue(storageDescriptor.getSerdeInfo().getParameters()
.containsKey(HiveMetadataPreservingTableOperations.ORC_COLUMNS_TYPES));

// Use same schema literal but containing uppercase and check no mismatch detected
tbl.setParameters(ImmutableMap.of("avro.schema.literal", testSchemaLiteralWithUppercase));
Assert.assertFalse(HiveMetadataPreservingTableOperations.fixMismatchedSchema(tbl));
}
}