ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

[SPARK-7330] [SQL] avoid NPE at jdbc rdd

Browse source 
Thank nadavoosh point this out in #5590

Author: Daoyuan Wang <daoyuan.wang@intel.com>

Closes #5877 from adrian-wang/jdbcrdd and squashes the following commits:

cc11900 [Daoyuan Wang] avoid NPE in jdbcrdd
This commit is contained in:
Daoyuan Wang 2015-05-07 10:05:01 -07:00 committed by Yin Huai
parent 4f87e9562a
commit ed9be06a47
2 changed files with 32 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
sql/core/src/main/scala/org/apache/spark/sql/jdbc/JDBCRDD.scala
View file

@ -362,7 +362,13 @@ private[sql] class JDBCRDD(
conversions(i) match {
case BooleanConversion => mutableRow.setBoolean(i, rs.getBoolean(pos))
case DateConversion =>
mutableRow.update(i, DateUtils.fromJavaDate(rs.getDate(pos)))
// DateUtils.fromJavaDate does not handle null value, so we need to check it.
val dateVal = rs.getDate(pos)
if (dateVal != null) {
mutableRow.update(i, DateUtils.fromJavaDate(dateVal))
} else {
mutableRow.update(i, null)
}
case DecimalConversion =>
val decimalVal = rs.getBigDecimal(pos)
if (decimalVal == null) {

25
sql/core/src/test/scala/org/apache/spark/sql/jdbc/JDBCSuite.scala
View file

@ -104,6 +104,8 @@ class JDBCSuite extends FunSuite with BeforeAndAfter {
).executeUpdate()
conn.prepareStatement("insert into test.timetypes values ('12:34:56', "
+ "'1996-01-01', '2002-02-20 11:22:33.543543543')").executeUpdate()
conn.prepareStatement("insert into test.timetypes values ('12:34:56', "
+ "null, '2002-02-20 11:22:33.543543543')").executeUpdate()
conn.commit()
sql(
s"""
@ -127,6 +129,23 @@ class JDBCSuite extends FunSuite with BeforeAndAfter {
|OPTIONS (url '$url', dbtable 'TEST.FLTTYPES', user 'testUser', password 'testPass')
""".stripMargin.replaceAll("\n", " "))
conn.prepareStatement(
s"""
|create table test.nulltypes (a INT, b BOOLEAN, c TINYINT, d BINARY(20), e VARCHAR(20),
|f VARCHAR_IGNORECASE(20), g CHAR(20), h BLOB, i CLOB, j TIME, k DATE, l TIMESTAMP,
|m DOUBLE, n REAL, o DECIMAL(40, 20))
""".stripMargin.replaceAll("\n", " ")).executeUpdate()
conn.prepareStatement("insert into test.nulltypes values ("
+ "null, null, null, null, null, null, null, null, null, "
+ "null, null, null, null, null, null)").executeUpdate()
conn.commit()
sql(
s"""
|CREATE TEMPORARY TABLE nulltypes
|USING org.apache.spark.sql.jdbc
|OPTIONS (url '$url', dbtable 'TEST.NULLTYPES', user 'testUser', password 'testPass')
""".stripMargin.replaceAll("\n", " "))
// Untested: IDENTITY, OTHER, UUID, ARRAY, and GEOMETRY types.
}
@ -254,6 +273,7 @@ class JDBCSuite extends FunSuite with BeforeAndAfter {
val rows = TestSQLContext.jdbc(urlWithUserAndPass, "TEST.TIMETYPES").collect()
val cachedRows = TestSQLContext.jdbc(urlWithUserAndPass, "TEST.TIMETYPES").cache().collect()
assert(rows(0).getAs[java.sql.Date](1) === java.sql.Date.valueOf("1996-01-01"))
assert(rows(1).getAs[java.sql.Date](1) === null)
assert(cachedRows(0).getAs[java.sql.Date](1) === java.sql.Date.valueOf("1996-01-01"))
}
@ -266,6 +286,11 @@ class JDBCSuite extends FunSuite with BeforeAndAfter {
assert(cachedRows(0).getAs[java.sql.Date](1) === java.sql.Date.valueOf("1996-01-01"))
}
test("test types for null value") {
val rows = TestSQLContext.jdbc(urlWithUserAndPass, "TEST.NULLTYPES").collect()
assert((0 to 14).forall(i => rows(0).isNullAt(i)))
}
test("H2 floating-point types") {
val rows = sql("SELECT * FROM flttypes").collect()
assert(rows(0).getDouble(0) === 1.00000000000000022) // Yes, I meant ==.