lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-916]: Added support for multiple input formats in TimestampBasedKeyGenerator (#1648)

Browse Source
This commit is contained in:
Pratyaksh Sharma
2020-07-11 00:58:45 +05:30
committed by GitHub
parent c7f1a781ab
commit 9627a385fe
7 changed files with 404 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

170
hudi-spark/src/test/java/org/apache/hudi/keygen/TestTimestampBasedKeyGenerator.java
View File

@@ -25,6 +25,8 @@ import org.apache.hudi.common.testutils.SchemaTestUtil;
import org.apache.avro.Schema;
import org.apache.avro.generic.GenericRecord;
import org.apache.hudi.exception.HoodieDeltaStreamerException;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
@@ -48,18 +50,41 @@ public class TestTimestampBasedKeyGenerator {
}
private TypedProperties getBaseKeyConfig(String timestampType, String dateFormat, String timezone, String scalarType) {
properties.setProperty("hoodie.deltastreamer.keygen.timebased.timestamp.type", timestampType);
properties.setProperty("hoodie.deltastreamer.keygen.timebased.output.dateformat", dateFormat);
properties.setProperty("hoodie.deltastreamer.keygen.timebased.timezone", timezone);
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_TYPE_FIELD_PROP, timestampType);
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_OUTPUT_DATE_FORMAT_PROP, dateFormat);
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_TIMEZONE_FORMAT_PROP, timezone);
if (scalarType != null) {
properties.setProperty("hoodie.deltastreamer.keygen.timebased.timestamp.scalar.time.unit", scalarType);
}
return properties;
}
private TypedProperties getBaseKeyConfig(String timestampType, String inputFormatList, String inputFormatDelimiterRegex, String inputTimezone, String outputFormat, String outputTimezone) {
if (timestampType != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_TYPE_FIELD_PROP, timestampType);
}
if (inputFormatList != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_INPUT_DATE_FORMAT_PROP, inputFormatList);
}
if (inputFormatDelimiterRegex != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_INPUT_DATE_FORMAT_LIST_DELIMITER_REGEX_PROP, inputFormatDelimiterRegex);
}
if (inputTimezone != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_INPUT_TIMEZONE_FORMAT_PROP, inputTimezone);
}
if (outputFormat != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_OUTPUT_DATE_FORMAT_PROP, outputFormat);
}
if (outputTimezone != null) {
properties.setProperty(TimestampBasedKeyGenerator.Config.TIMESTAMP_OUTPUT_TIMEZONE_FORMAT_PROP, outputTimezone);
}
return properties;
}
@Test
public void testTimestampBasedKeyGenerator() {
public void testTimestampBasedKeyGenerator() throws IOException {
// timezone is GMT+8:00
baseRecord.put("createTime", 1578283932000L);
properties = getBaseKeyConfig("EPOCHMILLISECONDS", "yyyy-MM-dd hh", "GMT+8:00", null);
@@ -85,7 +110,7 @@ public class TestTimestampBasedKeyGenerator {
}
@Test
public void testScalar() {
public void testScalar() throws IOException {
// timezone is GMT+8:00
baseRecord.put("createTime", 20000L);
@@ -94,4 +119,137 @@ public class TestTimestampBasedKeyGenerator {
HoodieKey hk5 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
assertEquals(hk5.getPartitionPath(), "2024-10-04 12");
}
}
@Test
public void test_ExpectsMatch_SingleInputFormat_ISO8601WithMsZ_OutputTimezoneAsUTC() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33.428Z");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"GMT");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040113", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_SingleInputFormats_ISO8601WithMsZ_OutputTimezoneAsInputDateTimeZone() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33.428Z");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040113", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ISO8601WithMsZ_OutputTimezoneAsUTC() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33.428Z");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"UTC");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040113", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ISO8601NoMsZ_OutputTimezoneAsUTC() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33Z");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"UTC");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040113", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ISO8601NoMsWithOffset_OutputTimezoneAsUTC() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33-05:00");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"UTC");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040118", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ISO8601WithMsWithOffset_OutputTimezoneAsUTC() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33.123-05:00");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"UTC");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040118", hk1.getPartitionPath());
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ISO8601WithMsZ_OutputTimezoneAsEST() throws IOException {
baseRecord.put("createTime", "2020-04-01T13:01:33.123Z");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"EST");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("2020040109", hk1.getPartitionPath());
}
@Test
public void test_Throws_MultipleInputFormats_InputDateNotMatchingFormats() {
baseRecord.put("createTime", "2020-04-01 13:01:33.123-05:00");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ",
"",
"",
"yyyyMMddHH",
"UTC");
Assertions.assertThrows(HoodieDeltaStreamerException.class, () -> new TimestampBasedKeyGenerator(properties).getKey(baseRecord));
}
@Test
public void test_ExpectsMatch_MultipleInputFormats_ShortDate_OutputCustomDate() throws IOException {
baseRecord.put("createTime", "20200401");
properties = this.getBaseKeyConfig(
"DATE_STRING",
"yyyy-MM-dd'T'HH:mm:ssZ,yyyy-MM-dd'T'HH:mm:ss.SSSZ,yyyyMMdd",
"",
"UTC",
"MM/dd/yyyy",
"UTC");
HoodieKey hk1 = new TimestampBasedKeyGenerator(properties).getKey(baseRecord);
Assertions.assertEquals("04/01/2020", hk1.getPartitionPath());
}
}