FrankChen021 commented on code in PR #18477:
URL: https://github.com/apache/druid/pull/18477#discussion_r2326702918
##########
processing/src/main/java/org/apache/druid/java/util/common/Intervals.java:
##########
@@ -53,6 +57,74 @@ public static Interval of(String format, Object...
formatArgs)
return of(StringUtils.format(format, formatArgs));
}
+ /**
+ * A performance-optimized method for parsing a Joda-Time {@link Interval}
from a string.
+ * This method is significantly faster than the standard {@link
Intervals#of(String)} for the following
+ * group of offsets:
+ * <ol>
+ * <li>"2022-01-01T00:00:00.000Z/2022-01-02T00:00:00.000Z"</li>
+ * <li>"2022-01-01T00:00:00.000+05:30/2022-01-01T01:00:00.000+05:30"</li>
+ * <li>"2022-01-01T00:00:00.000+0530/2022-01-01T01:00:00.000+0530"</li>
+ * </ol>
+ * <p>
+ * If the input string does not match the format, it will fall back to the
more flexible but
+ * slower {@link Intervals#of(String)} parser. If you are dealing with any
Intervals format examples below,
+ * consider using {@link Intervals#of(String)} instead:
+ * <ol>
+ * <li>"2022-01-01T00:00:00Z/2022-01-02T00:00:00Z" (without millis)</li>
+ * <li>"2022-01-01/2022-01-02" (Date only)</li>
+ * <li>"2022-01-01T12:00:00.000Z/PT6H" (Periods in start / end)</li>
+ * </ol>
+ *
+ * Currently, this method is only used in {@link
org.apache.druid.timeline.SegmentId}.
+ * Should you find additional places where the Interval format is guaranteed
to be compatible with this method,
+ * feel free to open a PR.
Review Comment:
Since this method is an optimization and used by the DataSegment only, my
opinion is to move this method from the Interval class to the DataSegment.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
