Skip to content

Commit

Permalink
[fix](Nereids) fix functions use split and add more cases about fe fo…
Browse files Browse the repository at this point in the history 
…ld string function
  • Loading branch information
@LiBinfeng-01
LiBinfeng-01 committed Dec 18, 2024
1 parent ed11252 commit b607a04
Show file tree
Hide file tree
Showing 2 changed files with 619 additions and 713 deletions.
55 changes: 39 additions & 16 deletions ...ava/org/apache/doris/nereids/trees/expressions/functions/executable/StringArithmetic.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -321,6 +321,19 @@ public static Expression locate(StringLikeLiteral first, StringLikeLiteral secon
return new IntegerLiteral(second.getValue().trim().indexOf(first.getValue()) + 1);
}

/**
* Executable arithmetic functions Locate
*/
@ExecFunction(name = "locate")
public static Expression locate(StringLikeLiteral first, StringLikeLiteral second, IntegerLiteral third) {
int result = second.getValue().trim().indexOf(first.getValue()) + 1;
if (third.getValue() <= 0 || !substringImpl(second.getValue(), third.getValue(),
second.getValue().length()).trim().contains(first.getValue())) {
result = 0;
}
return new IntegerLiteral(result);
}

/**
* Executable arithmetic functions Instr
*/
Expand All @@ -333,12 +346,14 @@ public static Expression instr(StringLikeLiteral first, StringLikeLiteral second
* Executable arithmetic functions Ascii
*/
@ExecFunction(name = "ascii")
public static Expression ascii(StringLikeLiteral first) {
public static Expression ascii(StringLikeLiteral first) throws UnsupportedEncodingException {
if (first.getValue().length() == 0) {
return new IntegerLiteral(0);
}
char firstChar = first.getValue().charAt(0);
return new IntegerLiteral(firstChar);
String character = first.getValue();
byte[] utf8Bytes = character.getBytes("UTF-8");
int firstByteAscii = utf8Bytes[0] & 0xFF;
return new IntegerLiteral(firstByteAscii);
}

/**
Expand Down Expand Up @@ -583,7 +598,7 @@ public static Expression fieldVarchar(StringLikeLiteral first, VarcharLiteral...
}

private static int findStringInSet(String target, String input) {
String[] split = input.split(",");
String[] split = input.split(",", -1);
for (int i = 0; i < split.length; i++) {
if (split[i].equals(target)) {
return i + 1;
Expand All @@ -605,6 +620,10 @@ public static Expression findInSetVarchar(StringLikeLiteral first, StringLikeLit
*/
@ExecFunction(name = "repeat")
public static Expression repeat(StringLikeLiteral first, IntegerLiteral second) {
// when it is too large for fe to make result string, do not folding on fe, limit 1 MB
if ((first.getValue().length() * second.getValue()) > 1000000) {
throw new AnalysisException("repeat too large to fold const by fe");
}
StringBuilder sb = new StringBuilder();
for (int i = 0; i < second.getValue(); i++) {
sb.append(first.getValue());
Expand All @@ -627,6 +646,10 @@ public static Expression reverseVarchar(StringLikeLiteral first) {
*/
@ExecFunction(name = "space")
public static Expression space(IntegerLiteral first) {
// when it is too large for fe to make result string, do not folding on fe, limit 1 MB
if (first.getValue() > 1000000) {
throw new AnalysisException("space too large to fold const by fe");
}
StringBuilder sb = new StringBuilder();
for (int i = 0; i < first.getValue(); i++) {
sb.append(' ');
Expand All @@ -635,13 +658,13 @@ public static Expression space(IntegerLiteral first) {
}

/**
* Executable arithmetic functions split_by_char
* Executable arithmetic functions split_by_string
*/
@ExecFunction(name = "split_by_char")
public static Expression splitByChar(StringLikeLiteral first, StringLikeLiteral second) {
String[] result = first.getValue().split(second.getValue());
@ExecFunction(name = "split_by_string")
public static Expression splitByString(StringLikeLiteral first, StringLikeLiteral second) {
String[] result = first.getValue().split(second.getValue(), -1);
List<Literal> items = new ArrayList<>();
for (int i = 1; i < result.length; i++) {
for (int i = 0; i < result.length; i++) {
items.add((Literal) castStringLikeLiteral(first, result[i]));
}
return new ArrayLiteral(items);
Expand All @@ -666,16 +689,16 @@ public static Expression splitPart(StringLikeLiteral first, StringLikeLiteral ch
if (".$|()[{^?*+\\".contains(separator) || separator.startsWith("\\")) {
separator = "\\" + separator;
}
parts = sb.reverse().toString().split(separator);
parts = sb.reverse().toString().split(separator, -1);
} else {
if (".$|()[{^?*+\\".contains(separator) || separator.startsWith("\\")) {
separator = "\\" + separator;
}
parts = first.getValue().split(separator);
parts = first.getValue().split(separator, -1);
}

if (parts.length < Math.abs(number.getValue()) || number.getValue() == 0) {
if (parts.length == Math.abs(number.getValue()) - 1) {
if (parts.length == Math.abs(number.getValue())) {
if (number.getValue() < 0 && first.getValue().startsWith(chr.getValue())
|| number.getValue() > 0 && first.getValue().endsWith(chr.getValue())) {
return castStringLikeLiteral(first, "");
Expand All @@ -695,7 +718,7 @@ public static Expression splitPart(StringLikeLiteral first, StringLikeLiteral ch
*/
@ExecFunction(name = "substring_index")
public static Expression substringIndex(StringLikeLiteral first, StringLikeLiteral chr, IntegerLiteral number) {
String[] parts = first.getValue().split(chr.getValue());
String[] parts = first.getValue().split(chr.getValue(), -1);
if (Math.abs(number.getValue()) >= parts.length) {
return first;
}
Expand Down Expand Up @@ -907,13 +930,13 @@ public static Expression extractUrlParameter(StringLikeLiteral first, StringLike
return castStringLikeLiteral(first, "");
}

String[] urlParts = first.getValue().split("\\?");
String[] urlParts = first.getValue().split("\\?", -1);
if (urlParts.length > 1) {
String query = urlParts[1];
String[] pairs = query.split("&");
String[] pairs = query.split("&", -1);

for (String pair : pairs) {
String[] keyValue = pair.split("=");
String[] keyValue = pair.split("=", -1);
if (second.getValue().equals(keyValue[0])) {
return castStringLikeLiteral(first, keyValue[1]);
}
Expand Down
Loading

0 comments on commit b607a04

Please sign in to comment.