Re: [Gnumed-bugs] [Gnumed-devel] 1.0.rc5 release

[Karsten Hilbert]

On Thu, Oct 06, 2011 at 10:04:02AM -0700, Jim Busser wrote:

> So as to unclutter the output, I constrained on pk > 2:
> 
>       select pk_audit, pk, name_short, version,
>       row_version, modified_when, modified_by
>       from ref.data_source where pk > 2 ;
> 
> gnumed_v15->
> 
>  pk_audit | pk | name_short |    version    | row_version |         
> modified_when         | modified_by 
> ----------+----+------------+---------------+-------------+-------------------------------+-------------
>     12016 |  3 | Aust NIPS  | 2005          |           0 | 2010-07-05 
> 11:51:09.288217-07 | gm-dbo
>     12111 |  4 | SFCVC      | December 2004 |           0 | 2010-07-05 
> 11:51:10.674081-07 | gm-dbo
>     12772 |  5 | ATC        | 2009-01-EN    |           0 | 2011-04-08 
> 12:38:58.583321-07 | gm-dbo
>     12773 |  6 | ATC        | 2009-01-DE    |           0 | 2011-04-08 
> 12:38:58.583321-07 | gm-dbo
>     12774 |  7 | ATC        | 2009-01-FR    |           0 | 2011-04-08 
> 12:38:58.583321-07 | gm-dbo
>     13331 |  8 | FreeDiams  | 0.5.4         |           0 | 2011-05-05 
> 14:39:50.368841-07 | any-doc
>     13891 | 10 | LOINC      | 2.26          |           0 | 2011-07-28 
> 08:13:24.483741-07 | any-doc
> (7 rows)
> 
> Which confirms the absence of pk=9, only I am not sure why.

This fosters the suspicion that the existing LOINC Update
facility isn't being quite as careful as it should.

> I then checked the audit table
> 
>       audit.log_data_source
> 
> as
> 
>       select pk_audit, pk, name_short, orig_when,
>       modified_when, orig_by, modified_by, audit_when, audit_by
>       from audit.log_data_source ;
> 
> and got
> 
> pk_audit | pk | name_short |           orig_when           |         
> modified_when         | orig_by | modified_by |          audit_when           
> | audit_by 
> ----------+----+------------+-------------------------------+-------------------------------+---------+-------------+-------------------------------+----------
>     13374 |  9 | LOINC      | 2011-05-10 00:03:48.570398-07 | 2011-05-10 
> 00:03:48.570398-07 | any-doc | any-doc     | 2011-07-28 08:13:24.483741-07 | 
> gm-dbo
> (1 row)
> 
> I cannot remember myself doing any primary import of LOINC so wonder whether 
> pk 9 might have gotten created using
> 
>       Master data > Update LOINC
> 
> and, since I cannot remember myself deleting pk 9 (though it is possible) 
> whether on July 28th it could have been the rerunning of
> 
>       Master data > Update LOINC
> 
> which deleted pk 9 as part of importing (creating) a new LOINC
> 
> ?

I suppose your suspicion is correct. Now, looking at

        client/business/gmLoinc.py::loinc_import()

confirms at least the possibility of the above being the reason.

(That file is, BTW, a good cursor on how to approach a LOINC
 data pack ;-)

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346