Summary: When performing a search and sorting by date, the sorting appears to be incorrect.
- The date associated with the indexed documents may be missing or incorrect.
- Sorting parameter may not have been specified correctly.
- Verify correct dates are associated with the indexed documents by checking the <FS> tag values. In case dates in <FS> tags are missing or incorrect, double check settings in Crawl and Index > Document Dates. See Document Dates Help Page for more details. Instructions on how to check the <FS> tag are described below in the "Additional Details" section.
- Make sure proper date sort parameter values are used. For details see the Sort by Date document.
- To see the <FS> tag values for indexed documents, do a search against the collection with XML output, for example, "http://GSA/search?q=<search term>&site=default_collection&getfields=*". In the XML output check for the <FS> tag which should be in this format:
<FS name="date" value="YYYY-MM-DD">
Incorrect document dates settings might result in this value being empty or incorrect.
- Note that in case you have multiple rules configured in Document Dates, for any given URL the first rule whose pattern matches the URL will be picked. So you need to put more restrictive patterns first, with the catch-all pattern at the bottom.
- Also note that, it is possible to define date format, locale and prefix string. See above mentioned Document Dates Help Page for more details. One thing to note is that format should be what is required for the strptime Linux system call, so, if the there is a meta tag such as this one: <meta name='pubdate' content='20120223'> in web page, then the format string needs to be %Y%m%d (not "YYYYMMDD").
- Changes to Document Dates rules will become effective when the affected URLs are re-crawled.