Bug #2651
closed
select="xs:string(xs:dateTime('9999-12-31T23:59:59.9999999'))" gives => '9999-12-31T23:59:59.:"
Category:
XPath conformance
Applies to branch:
9.6, 9.7
Fix Committed on Branch:
9.6, 9.7
Fixed in Maintenance Release:
Description
Running the following xslq;
<xsl:stylesheet version="2.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:xs="http://www.w3.org/2001/XMLSchema" exclude-result-prefixes="xs">
<xsl:output method="xml" encoding="UTF-8" byte-order-mark="no" indent="yes"/>
<xsl:template match="/">
<xsl:sequence select="xs:string(xs:dateTime('9999-12-31T23:59:59.9999999'))"/>
</xsl:template>
</xsl:stylesheet>
Results in;
9999-12-31T23:59:59.:
Notice the : at the end of the string
If we remove the last 9 from the datetime or change it to a value below 9 results in the following;
9999-12-31T23:59:59.999999
Cheers
Stew
- Category set to XPath conformance
- Assignee set to Debbie Lockett
I have added this to the QT3 test suite as CastExpr/CastAs676
- Fix Committed on Branch 9.6, 9.7 added
Bug fix committed on 9.6 and 9.7 branches.
Fix in DateTimeValue.makeDateTimeValue: if the microseconds value rounds to the upper bound of 1,000,000, then truncate to 999,999.
(So in QT3 test CastExpr/CastAs676 we now return 9999-12-31T23:59:59.999999)
- Status changed from New to Resolved
Why is there a limit of 6 digits in the factional part of a second? I thought that the W3 standard was for any number of digits?
The max DateTime value in dot net is 9999-12-31T23:59:59.9999999 (a precision of 100 nanoseconds), if you truncate this to 9999-12-31T23:59:59.999999 (a precision of 1 microsecond) you cannot test for value == DateTime.max.
The W3C specs require support for millisecond precision (3 fractional digits) in xs:dateTime (See Functions and Operators section 9.1.1). Saxon chooses to support microsecond precision (6 digits). I've no particular recollection about why this figure was chosen.
As far as I can see, there's nothing in the specs that says what you should do if casting from a string with more digits than this, but analogy with other data types suggests that rounding is a reasonable strategy. However, when we thought about your edge case, we decided that rounding the milliseconds in such a way that the date changes would not be what users wanted; if they are using a value like this, they almost certainly intend it to mean "the last representable point in time before the next day starts", so truncation is more appropriate. The alternative strategy, of rounding to 10000-01-01T00:00:00 would also create a value that's not equal to .NET's DateTime.max.
Michael
Thanks for you prompt response, the level of support here is second to none, awesome job.
I agree with your reasoning here with regard to truncating as apposed to rounding but given that the native DateTime in dot net uses 100 nanosecond or 7 digit precision it would be best if the dot net (or all) version(s) of Saxon used 7 digits for DateTime values.
- Applies to branch 9.7 added
- % Done changed from 0 to 100
- Fixed in Maintenance Release 9.7.0.4 added
Bug fix applied in the Saxon 9.7.0.4 maintenance release. Leave open until patched applied in the 9.6 maintenance release.
- Status changed from Resolved to Closed
- Fixed in Maintenance Release 9.6.0.9 added
- Fixed in Maintenance Release deleted (
9.7.0.4)
Bug fix applied in the Saxon 9.6.0.9 maintenance release.
- Sprint/Milestone set to 9.7.0.4
- Fixed in Maintenance Release 9.7.0.4 added
Please register to edit this issue
Also available in: Atom
PDF