Closed csaundrs closed 8 years ago
On the 3 that have bib_id ranges, should I take out that condition altogether or should I hard code the upper and lower limits to certain values?
Since we have databases of varying sizes, I think it would be simplest if you can take out the condition altogether so that any library running this report will find all of the matching bibs in their entire database. Unless you see something that makes this a bad idea for performance or unreasonable results times, of course!
Serial bib records without an ISSN (Local Cat Maint > Bibliographic Records: Control Numbers) --In KNXdb on devel, web-reports is returning zero results, but in Access this query returns 1873 results.
There were two problems with the way this report was entered. The first was easy to fix. BIB_TEXT did not belong before the column names and also the alias (or whatever it's called in Oracle world) for "SUBSTR(FIELD_008,7,1)" Publication Status should have been entered instead:
About the second problem: some googling revealed that Oracle treats empty strings as null. So I replaced ' ' with null. Now it returns 1873 rows in web-reports, same as @gibsonjc reported in Access.
I'm troubleshooting
Find 020's with multiple $'s (Local Cat Maint > Bibliographic Records: Control Numbers) --In KNXdb on devel, web-reports is returning zero results, but in Access this query returns 2924 results.
Jessica, Can you please post some matches for this query from Access or (preferably) email me the Access file if it's not too large (if needed you can right click on the file in macOS and click compress...). That would be a great help.
Thanks.
This is about "Find 020's with multiple $'s (Local Cat Maint > Bibliographic Records: Control Numbers)."
Chris, did you craft the regexp yourself used in the query or did you get it from somewhere else? I'm no expert on regular expressions by any means but I think I see a few problems with it (for one, I don't think the use of anchors belongs?). Unfortunately, my attempt suffers from a problem too. With my mine I get 2,925 results compared to Jessica's 2,924. I used ([0-9]{9}).*{2,}
I don't think this one is truly correct since it will only match on double ISBNs, not entries with 3 or more.
Oops, I copy pasted the wrong regexp that I used. I meant to paste. '([0-9]{9}).*([0-9]{9})')
I'm making typos left and right. In my last post I left an errant parenthesis on the right end of the string that shouldn't have been there.
Anyway, I think I may have come up with one that should match 2 or more ISBNs: ([0-9]{9}).*([0-9]{9}).*{1,}
It still returns the same amount for KNX: 2,925, but Jessica, could that be because there simply aren't any 020 fields that contain more than 2 ISBNs in KNXdb? I ran the query on UIU and found some results like:
8470900315 (set) 8470900323 (v.1) 847090048X (v.2) 3487041766 (v.1) 3487041774 (v.2) 3487041782 (v.3)
Just sent you the KNXdb on devel Access output via email (Excel spreadsheet).
KNXdb does have bibs with more than two 020 $a's like this one that can be found in both the Access and the web-reports output: BIB ID 1689 0873952855 0873952863 (pbk.) 0873952871 (micro.)
KNXdb does have bibs with more than two 020 $a's like this one that can be found in both the Access and the web-reports output:
I was confused why my first regex that I posted would return bibs with more than two ISBNs when I expected it to match two and only two ISBNs. But I made a really dumb mistake in my regexp: ".*
"
The error is obviously present in([0-9]{9}).*([0-9]{9}).*{1,}
too.
I'll work on refining it further.
Closing. Split apart into 102, 103, and 104.
If these check out ok, they are ready for production: ( I found that one of them, designated by \ does not work, but the rest return either results or zero rows)