[
https://issues.apache.org/jira/browse/SQOOP-2906?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15237015#comment-15237015
]
ASF GitHub Bot commented on SQOOP-2906:
---------------------------------------
Github user stanleyxu2005 commented on a diff in the pull request:
https://github.com/apache/sqoop/pull/18#discussion_r59360059
--- Diff: src/java/org/apache/sqoop/avro/AvroUtil.java ---
@@ -114,11 +114,20 @@ public static String toAvroColumn(String column) {
* Format candidate to avro specifics
*/
public static String toAvroIdentifier(String candidate) {
- String formattedCandidate = candidate.replaceAll("\\W+", "_");
- if (formattedCandidate.substring(0,1).matches("[a-zA-Z_]")) {
- return formattedCandidate;
+ char[] data = candidate.toCharArray();
+ int stringIndex = 0;
+
+ for (char c:data) {
+ if (Character.isLetterOrDigit(c) || c == '_') {
+ data[stringIndex++] = c;
+ }
+ }
+
+ char initial = data[0];
+ if (Character.isLetter(initial) || initial == '_') {
+ return new String(data, 0, stringIndex);
--- End diff --
Your code will first create a char array and then eventually update char in
the array. As result you will create another copy as a new String. Have you
thought about using a `StringBuilder` directly?
```
final StringBuilder sb = new StringBuilder();
for (char c : candidate) {
if (Character.isLetterOrDigit(c) || c == '_') {
sb.append(c);
}
}
...
return sb.toString();
```
> Optimization of AvroUtil.toAvroIdentifier
> -----------------------------------------
>
> Key: SQOOP-2906
> URL: https://issues.apache.org/jira/browse/SQOOP-2906
> Project: Sqoop
> Issue Type: Improvement
> Reporter: Joeri Hermans
> Labels: avro, hadoop, optimization
>
> Hi all
> Our distributed profiler indicated some inefficiencies in the
> AvroUtil.toAvroIdentifier method, more specifically, the use of Regex
> patterns. This can be directly observed from the FlameGraph generated by this
> profiler (https://jhermans.web.cern.ch/jhermans/sqoop_avro_flamegraph.svg).
> We implemented an optimization, and compared this with the original method.
> On our testing machine, the optimization by itself is about 500% (on average)
> more efficient compared to the original implementation. We have yet to test
> how this optimization will influence the performance of user jobs.
> Any suggestions or remarks are welcome.
> Kind regards,
> Joeri
> https://github.com/apache/sqoop/pull/18
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
