[
https://issues.apache.org/jira/browse/DRILL-5450?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15992350#comment-15992350
]
ASF GitHub Bot commented on DRILL-5450:
---------------------------------------
Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/821#discussion_r114249116
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/expr/fn/impl/StringFunctionHelpers.java
---
@@ -144,41 +144,28 @@ public static int varTypesToInt(final int start,
final int end, DrillBuf buffer)
return result;
}
- // Assumes Alpha as [A-Za-z0-9]
- // white space is treated as everything else.
+ /**
+ * Capitalizes first letter in each word.
+ * Any symbol except digits and letters is considered as word delimiter.
+ *
+ * @param start start position in input buffer
+ * @param end end position in input buffer
+ * @param inBuf buffer with input characters
+ * @param outBuf buffer with output characters
+ */
public static void initCap(int start, int end, DrillBuf inBuf, DrillBuf
outBuf) {
- boolean capNext = true;
+ boolean capitalizeNext = true;
int out = 0;
for (int id = start; id < end; id++, out++) {
- byte currentByte = inBuf.getByte(id);
-
- // 'A - Z' : 0x41 - 0x5A
- // 'a - z' : 0x61 - 0x7A
- // '0-9' : 0x30 - 0x39
- if (capNext) { // curCh is whitespace or first character of word.
- if (currentByte >= 0x30 && currentByte <= 0x39) { // 0-9
- capNext = false;
- } else if (currentByte >= 0x41 && currentByte <= 0x5A) { // A-Z
- capNext = false;
- } else if (currentByte >= 0x61 && currentByte <= 0x7A) { // a-z
- capNext = false;
- currentByte -= 0x20; // Uppercase this character
- }
- // else {} whitespace
- } else { // Inside of a word or white space after end of word.
- if (currentByte >= 0x30 && currentByte <= 0x39) { // 0-9
- // noop
- } else if (currentByte >= 0x41 && currentByte <= 0x5A) { // A-Z
- currentByte -= 0x20; // Lowercase this character
- } else if (currentByte >= 0x61 && currentByte <= 0x7A) { // a-z
- // noop
- } else { // whitespace
- capNext = true;
- }
+ int currentByte = inBuf.getByte(id);
--- End diff --
This code works only for ASCII, but not for UTF-8. UTF-8 is a multi-byte
code that requires special encoding/decoding to convert to Unicode characters.
Without that encoding, this method won't work for Cyrillic, Greek or any other
character set with upper/lower distinctions.
Since this method never worked, it is probably OK to make it a bit less
broken than before: at least now it works for ASCII. Please add unit tests
below, then file a JIRA, for the fact that this function does not work with
UTF-8 despite the fact that Drill claims it supports UTF-8.
> Fix initcap function to convert upper case characters correctly
> ---------------------------------------------------------------
>
> Key: DRILL-5450
> URL: https://issues.apache.org/jira/browse/DRILL-5450
> Project: Apache Drill
> Issue Type: Bug
> Components: Functions - Drill
> Affects Versions: 1.10.0
> Reporter: Arina Ielchiieva
> Assignee: Arina Ielchiieva
>
> Initcap function converts incorrectly subsequent upper case characters after
> first character.
> {noformat}
> 0: jdbc:drill:zk=local> select initcap('aaa') from (values(1));
> +---------+
> | EXPR$0 |
> +---------+
> | Aaa |
> +---------+
> 1 row selected (0.275 seconds)
> 0: jdbc:drill:zk=local> select initcap('AAA') from (values(1));
> +---------+
> | EXPR$0 |
> +---------+
> | A!! |
> +---------+
> 1 row selected (0.27 seconds)
> 0: jdbc:drill:zk=local> select initcap('aAa') from (values(1));
> +---------+
> | EXPR$0 |
> +---------+
> | A!a |
> +---------+
> 1 row selected (0.229 seconds)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
