Skip to content

[SPARK-48435][SQL] UNICODE collation should not support binary equality #46772

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

Closed
uros-db wants to merge 9 commits into from
Closed

[SPARK-48435][SQL] UNICODE collation should not support binary equality #46772

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
c0737f1
Initial commit
uros-db May 28, 2024
1e6bee0
Use TERTIARY strength
uros-db May 28, 2024
bfd7e60
Fix tests
uros-db May 28, 2024
319d0a8
Fix tests
uros-db May 28, 2024
b5501e8
Merge branch 'master' into fix-unicode
uros-db May 28, 2024
26fc209
Update CollationFactorySuite.scala
uros-db May 29, 2024
d0b741c
Merge branch 'master' into fix-unicode
uros-db May 29, 2024
6bf6e4e
Merge branch 'apache:master' into fix-unicode
uros-db May 30, 2024
fcf122a
Fix tests
uros-db May 30, 2024
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
12 changes: 9 additions & 3 deletions common/unsafe/src/main/java/org/apache/spark/sql/catalyst/util/CollationAwareUTF8String.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,13 @@
* Utility class for collation-aware UTF8String operations.
*/
public class CollationAwareUTF8String {

/**
* The constant value to indicate that the match is not found when searching for a pattern
* string in a target string.
*/
private static final int MATCH_NOT_FOUND = -1;

public static UTF8String replace(final UTF8String src, final UTF8String search,
final UTF8String replace, final int collationId) {
// This collation aware implementation is based on existing implementation on UTF8String
Expand Down Expand Up @@ -185,9 +192,8 @@ public static int findInSet(final UTF8String match, final UTF8String set, int co

public static int indexOf(final UTF8String target, final UTF8String pattern,
final int start, final int collationId) {
if (pattern.numBytes() == 0) {
return 0;
}
if (pattern.numBytes() == 0) return 0;
if (target.numBytes() == 0) return MATCH_NOT_FOUND;

StringSearch stringSearch = CollationFactory.getStringSearch(target, pattern, collationId);
stringSearch.setIndex(start);
Expand Down
4 changes: 2 additions & 2 deletions common/unsafe/src/main/java/org/apache/spark/sql/catalyst/util/CollationFactory.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -245,12 +245,12 @@ public CollationIdentifier identifier() {
PROVIDER_ICU,
Collator.getInstance(ULocale.ROOT),
"153.120.0.0",
true,
false,
Copy link
Contributor Author

@uros-db uros-db May 28, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

note for reviewers: this is the place we say UNICODE does not supportBinaryEquality

false,
false
);

collationTable[2].collator.setStrength(Collator.TERTIARY);
collationTable[2].collator.setStrength(Collator.IDENTICAL);
collationTable[2].collator.freeze();

// UNICODE case-insensitive comparison (ROOT locale, in ICU + Secondary strength).
Expand Down
36 changes: 2 additions & 34 deletions common/unsafe/src/test/java/org/apache/spark/unsafe/types/CollationSupportSuite.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -563,6 +563,8 @@ public void testStringInstr() throws SparkException {
assertStringInstr("", "xxxx", "UNICODE", 0);
assertStringInstr("test大千世界X大千世界", "界x", "UNICODE", 0);
assertStringInstr("test大千世界X大千世界", "界X", "UNICODE", 8);
assertStringInstr("xxxx", "", "UNICODE_CI", 1);
assertStringInstr("", "xxxx", "UNICODE_CI", 0);
assertStringInstr("aaads", "AD", "UNICODE_CI", 3);
assertStringInstr("aaads", "dS", "UNICODE_CI", 4);
assertStringInstr("test大千世界X大千世界", "界y", "UNICODE_CI", 0);
Expand Down Expand Up @@ -914,20 +916,6 @@ public void testStringTrim() throws SparkException {
assertStringTrim("UTF8_BINARY_LCASE", "xxasdxx", "x", "asd");
assertStringTrim("UTF8_BINARY_LCASE", "xa世ax", "x", "a世a");

assertStringTrimLeft("UNICODE", "asd", null, "asd");
assertStringTrimLeft("UNICODE", " asd ", null, "asd ");
assertStringTrimLeft("UNICODE", " a世a ", null, "a世a ");
assertStringTrimLeft("UNICODE", "asd", "x", "asd");
assertStringTrimLeft("UNICODE", "xxasdxx", "x", "asdxx");
assertStringTrimLeft("UNICODE", "xa世ax", "x", "a世ax");

assertStringTrimRight("UNICODE", "asd", null, "asd");
assertStringTrimRight("UNICODE", " asd ", null, " asd");
assertStringTrimRight("UNICODE", " a世a ", null, " a世a");
assertStringTrimRight("UNICODE", "asd", "x", "asd");
assertStringTrimRight("UNICODE", "xxasdxx", "x", "xxasd");
assertStringTrimRight("UNICODE", "xa世ax", "x", "xa世a");

// Test cases where trimString has more than one character
assertStringTrim("UTF8_BINARY", "ddsXXXaa", "asd", "XXX");
assertStringTrimLeft("UTF8_BINARY", "ddsXXXaa", "asd", "XXXaa");
Expand All @@ -937,22 +925,14 @@ public void testStringTrim() throws SparkException {
assertStringTrimLeft("UTF8_BINARY_LCASE", "ddsXXXaa", "asd", "XXXaa");
assertStringTrimRight("UTF8_BINARY_LCASE", "ddsXXXaa", "asd", "ddsXXX");

assertStringTrim("UNICODE", "ddsXXXaa", "asd", "XXX");
assertStringTrimLeft("UNICODE", "ddsXXXaa", "asd", "XXXaa");
assertStringTrimRight("UNICODE", "ddsXXXaa", "asd", "ddsXXX");

// Test cases specific to collation type
// uppercase trim, lowercase src
assertStringTrim("UTF8_BINARY", "asd", "A", "asd");
assertStringTrim("UTF8_BINARY_LCASE", "asd", "A", "sd");
assertStringTrim("UNICODE", "asd", "A", "asd");
assertStringTrim("UNICODE_CI", "asd", "A", "sd");

// lowercase trim, uppercase src
assertStringTrim("UTF8_BINARY", "ASD", "a", "ASD");
assertStringTrim("UTF8_BINARY_LCASE", "ASD", "a", "SD");
assertStringTrim("UNICODE", "ASD", "a", "ASD");
assertStringTrim("UNICODE_CI", "ASD", "a", "SD");

// uppercase and lowercase chars of different byte-length (utf8)
assertStringTrim("UTF8_BINARY", "ẞaaaẞ", "ß", "ẞaaaẞ");
Expand All @@ -963,10 +943,6 @@ public void testStringTrim() throws SparkException {
assertStringTrimLeft("UTF8_BINARY_LCASE", "ẞaaaẞ", "ß", "aaaẞ");
assertStringTrimRight("UTF8_BINARY_LCASE", "ẞaaaẞ", "ß", "ẞaaa");

assertStringTrim("UNICODE", "ẞaaaẞ", "ß", "ẞaaaẞ");
assertStringTrimLeft("UNICODE", "ẞaaaẞ", "ß", "ẞaaaẞ");
assertStringTrimRight("UNICODE", "ẞaaaẞ", "ß", "ẞaaaẞ");

assertStringTrim("UTF8_BINARY", "ßaaaß", "ẞ", "ßaaaß");
assertStringTrimLeft("UTF8_BINARY", "ßaaaß", "ẞ", "ßaaaß");
assertStringTrimRight("UTF8_BINARY", "ßaaaß", "ẞ", "ßaaaß");
Expand All @@ -975,10 +951,6 @@ public void testStringTrim() throws SparkException {
assertStringTrimLeft("UTF8_BINARY_LCASE", "ßaaaß", "ẞ", "aaaß");
assertStringTrimRight("UTF8_BINARY_LCASE", "ßaaaß", "ẞ", "ßaaa");

assertStringTrim("UNICODE", "ßaaaß", "ẞ", "ßaaaß");
assertStringTrimLeft("UNICODE", "ßaaaß", "ẞ", "ßaaaß");
assertStringTrimRight("UNICODE", "ßaaaß", "ẞ", "ßaaaß");

// different byte-length (utf8) chars trimmed
assertStringTrim("UTF8_BINARY", "Ëaaaẞ", "Ëẞ", "aaa");
assertStringTrimLeft("UTF8_BINARY", "Ëaaaẞ", "Ëẞ", "aaaẞ");
Expand All @@ -987,10 +959,6 @@ public void testStringTrim() throws SparkException {
assertStringTrim("UTF8_BINARY_LCASE", "Ëaaaẞ", "Ëẞ", "aaa");
assertStringTrimLeft("UTF8_BINARY_LCASE", "Ëaaaẞ", "Ëẞ", "aaaẞ");
assertStringTrimRight("UTF8_BINARY_LCASE", "Ëaaaẞ", "Ëẞ", "Ëaaa");

assertStringTrim("UNICODE", "Ëaaaẞ", "Ëẞ", "aaa");
assertStringTrimLeft("UNICODE", "Ëaaaẞ", "Ëẞ", "aaaẞ");
assertStringTrimRight("UNICODE", "Ëaaaẞ", "Ëẞ", "Ëaaa");
}

// TODO: Test more collation-aware string expressions.
Expand Down
7 changes: 6 additions & 1 deletion common/unsafe/src/test/scala/org/apache/spark/unsafe/types/CollationFactorySuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,18 +64,23 @@ class CollationFactorySuite extends AnyFunSuite with Matchers { // scalastyle:ig
CollationTestCase("UTF8_BINARY", "aaa", "aaa", true),
CollationTestCase("UTF8_BINARY", "aaa", "AAA", false),
CollationTestCase("UTF8_BINARY", "aaa", "bbb", false),
CollationTestCase("UTF8_BINARY", "å", "a\u030A", false),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "aaa", true),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "AAA", true),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "AaA", true),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "AaA", true),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "aa", false),
CollationTestCase("UTF8_BINARY_LCASE", "aaa", "bbb", false),
CollationTestCase("UTF8_BINARY_LCASE", "å", "a\u030A", false),
CollationTestCase("UNICODE", "aaa", "aaa", true),
CollationTestCase("UNICODE", "aaa", "AAA", false),
CollationTestCase("UNICODE", "aaa", "bbb", false),
CollationTestCase("UNICODE", "å", "a\u030A", true),
CollationTestCase("UNICODE_CI", "aaa", "aaa", true),
CollationTestCase("UNICODE_CI", "aaa", "AAA", true),
CollationTestCase("UNICODE_CI", "aaa", "bbb", false))
CollationTestCase("UNICODE_CI", "aaa", "bbb", false),
CollationTestCase("UNICODE_CI", "å", "a\u030A", true)
)

checks.foreach(testCase => {
val collation = fetchCollation(testCase.collationName)
Expand Down
32 changes: 17 additions & 15 deletions sql/core/src/test/scala/org/apache/spark/sql/CollationStringExpressionsSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -804,10 +804,12 @@ class CollationStringExpressionsSuite

test("StringTrim* functions - unit tests for both paths (codegen and eval)") {
// Without trimString param.
checkEvaluation(StringTrim(Literal.create( " asd ", StringType("UTF8_BINARY"))), "asd")
checkEvaluation(
StringTrim(Literal.create( " asd ", StringType("UTF8_BINARY"))), "asd")
checkEvaluation(
StringTrimLeft(Literal.create(" asd ", StringType("UTF8_BINARY_LCASE"))), "asd ")
checkEvaluation(StringTrimRight(Literal.create(" asd ", StringType("UNICODE"))), " asd")
checkEvaluation(StringTrimRight(
Literal.create(" asd ", StringType("UTF8_BINARY"))), " asd")

// With trimString param.
checkEvaluation(
Expand All @@ -822,8 +824,8 @@ class CollationStringExpressionsSuite
"asd ")
checkEvaluation(
StringTrimRight(
Literal.create(" asd ", StringType("UNICODE")),
Literal.create(" ", StringType("UNICODE"))),
Literal.create(" asd ", StringType("UTF8_BINARY")),
Literal.create(" ", StringType("UTF8_BINARY"))),
" asd")

checkEvaluation(
Expand All @@ -838,8 +840,8 @@ class CollationStringExpressionsSuite
"asdxx")
checkEvaluation(
StringTrimRight(
Literal.create("xxasdxx", StringType("UNICODE")),
Literal.create("x", StringType("UNICODE"))),
Literal.create("xxasdxx", StringType("UTF8_BINARY")),
Literal.create("x", StringType("UTF8_BINARY"))),
"xxasd")
}

Expand All @@ -863,10 +865,10 @@ class CollationStringExpressionsSuite
StringTrimTestCase("UTF8_BINARY_LCASE", "LTRIM", "xxasdxx", true, "x", "asdxx"),
StringTrimTestCase("UTF8_BINARY_LCASE", "RTRIM", " asd ", false, null, " asd"),

StringTrimTestCase("UNICODE", "TRIM", "xxasdxx", true, "x", "asd"),
StringTrimTestCase("UNICODE", "BTRIM", "xxasdxx", true, "x", "asd"),
StringTrimTestCase("UNICODE", "LTRIM", " asd ", false, null, "asd "),
StringTrimTestCase("UNICODE", "RTRIM", " asd ", true, null, null)
StringTrimTestCase("UTF8_BINARY", "TRIM", "xxasdxx", true, "x", "asd"),
StringTrimTestCase("UTF8_BINARY", "BTRIM", "xxasdxx", true, "x", "asd"),
StringTrimTestCase("UTF8_BINARY", "LTRIM", " asd ", false, null, "asd "),
StringTrimTestCase("UTF8_BINARY", "RTRIM", " asd ", true, null, null)

// Other more complex cases can be found in unit tests in CollationSupportSuite.java.
)
Expand Down Expand Up @@ -906,7 +908,7 @@ class CollationStringExpressionsSuite
+ "COLLATE('x', 'UTF8_BINARY_LCASE'))"),
expectedAnswer = Row("a"))
checkAnswer(
df = sql("SELECT LTRIM(COLLATE('x', 'UNICODE'), COLLATE('xax', 'UNICODE'))"),
df = sql("SELECT LTRIM(COLLATE('x', 'UTF8_BINARY'), COLLATE('xax', 'UTF8_BINARY'))"),
expectedAnswer = Row("ax"))

checkAnswer(
Expand All @@ -916,7 +918,7 @@ class CollationStringExpressionsSuite
df = sql("SELECT TRIM('x', COLLATE('xax', 'UTF8_BINARY_LCASE'))"),
expectedAnswer = Row("a"))
checkAnswer(
df = sql("SELECT BTRIM('xax', COLLATE('x', 'UNICODE'))"),
df = sql("SELECT BTRIM('xax', COLLATE('x', 'UTF8_BINARY'))"),
expectedAnswer = Row("a"))

checkAnswer(
Expand All @@ -926,21 +928,21 @@ class CollationStringExpressionsSuite
df = sql("SELECT RTRIM(COLLATE('x', 'UTF8_BINARY_LCASE'), 'xax')"),
expectedAnswer = Row("xa"))
checkAnswer(
df = sql("SELECT TRIM(COLLATE('x', 'UNICODE'), 'xax')"),
df = sql("SELECT TRIM(COLLATE('x', 'UTF8_BINARY'), 'xax')"),
expectedAnswer = Row("a"))
}

test("StringTrim* functions - collation type mismatch") {
List("TRIM", "LTRIM", "RTRIM").foreach(func => {
val collationMismatch = intercept[AnalysisException] {
sql("SELECT " + func + "(COLLATE('x', 'UTF8_BINARY_LCASE'), "
+ "COLLATE('xxaaaxx', 'UNICODE'))")
+ "COLLATE('xxaaaxx', 'UTF8_BINARY'))")
}
assert(collationMismatch.getErrorClass === "COLLATION_MISMATCH.EXPLICIT")
})

val collationMismatch = intercept[AnalysisException] {
sql("SELECT BTRIM(COLLATE('xxaaaxx', 'UNICODE'), COLLATE('x', 'UTF8_BINARY_LCASE'))")
sql("SELECT BTRIM(COLLATE('xxaaaxx', 'UTF8_BINARY'), COLLATE('x', 'UTF8_BINARY_LCASE'))")
}
assert(collationMismatch.getErrorClass === "COLLATION_MISMATCH.EXPLICIT")
}
Expand Down