IRC log for #koha, 2005-03-17

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
11:04 kados hi all
11:04 paul: I'm taking a look at searching queries today and I'm wondering what's the difference between these:
11:05 select distinct m1.bibid from marc_word as m1,marc_word as m2 where (m1.bibid=m2.bibid)and ((m1.word like 'block') and (m2.word like 'john'));
11:06 select distinct m1.bibid from biblio,biblioitems,marc_biblio,marc_word as m1,marc_word as m2 where biblio.biblionumber=marc_biblio.biblionumber and biblio.biblionumber=biblioitems.biblionumber and m1.bibid=marc_biblio.bibid and (m1.bibid=m2.bibid) and ((m1.word  like 'block')and (m2.word like 'john'))
11:06 (in my tests, with NPL data they seem to be the same ... am I missing something?)
11:07 paul you're right, they are the same. Except when you add "order by biblioitems.isbn" for example
11:07 ('morning kados)
11:07 kados (evening paul ;-))
11:08 ahh ... I see ... but I think the orderby is destroyed anyway
11:08 paul destroyed by who ?
11:08 NPL specific hack ?
11:09 kados I thought it was destroyed when the object was hashed
11:11 I think I'm wrong
11:11 so I had a real nice sort process that really messes up perfs for searches with result sets
11:12 it uses marc non-filing chars
11:12 I'm thinking of adding a colum to marc_word for biblio.title
11:12 (without the non-filing chars)
11:12 strictly for sorting purposes
11:13 so I can sort properly for English results (i.e., don't order by the articles a and the, etc.)
11:13 does that sound like the best solution?
11:14 it's too bad we can't make marc_word automatically sorted correctly
11:15 that would maybe be another way to do it
11:27 paul kados...
11:27 the marc_word is huge enough not to have another column !
11:28 i will add in the future (2 months ? 3 ?) a hack to order correctly the titles.
11:28 it will require a new column in biblio (let imagine biblio.Stitle, S being for "simplified")
11:29 the Stitle being filled with articles removed at the beginning of the field.
11:29 Also note that we should not expect too many perf improvements until we have marc_words.
11:30 I made some tests on my biggest DB (45000 biblios, 2 950 000 lines in marc_word)
11:30 with the marc_words, you have just 450 000 lines.
11:30 You can test yourself the effect for NPL with the following query :
11:31 select count(*), subfieldvalue,tag,subfieldcode from marc_word group by tag,subfieldcode,subfieldvalue
11:31 & count how many lines you will have.
11:31 I have some othe ideas to improve them more.
11:32 like "removing word indexes" option in MARC setup. To avoid indexing useless fields/subfields (like the 1xx ones in UNIMARC, or some others)
11:37 oups, my query does not work :
11:37 select count(*), word,tagsubfield from marc_word group by tagsubfield,word
11:37 is better
11:38 kados yea ... I've done that in the past and my result are similar
11:38 it will really speed up with marc_words
11:40 paul do you have the book "high performance mysql"?
11:40 paul nope
11:40 do you think I should buy it ?
11:40 kados yes
11:41 http://www.oreilly.com/catalog/hpmysql/
11:42 (unless you already know all that stuff ;-))
11:42 it specifically addresses application design
11:42 citing useing mysql in environments where the tables are several _billion_ rows
11:44 paul ok, i'm buying it now ;-) (even if it's in english only...)
11:57 hdl paul: fais moi signe quand tu seras disponible
11:58 paul ok, faut que je reboote, puis je te rappelle (je finis juste un mail)
11:58 hdl ok.
12:06 kados paul: i just sent you an email about marc non-filing chars in MARC21
12:13 paul: i just sent you an email about marc non-filing chars in MARC21 (it may be useful for your upcoming sort feature)
12:13 paul yes, it's in my mailbox now
17:07 rach morning
17:09 kados hi rach
17:11 rach how's the new site coming joshua?
17:37 hdl hi rach
18:11 kados rach: it's coming ;-) we've launched it at liblime.com but are waiting until the demos are done (and priv/terms) before we make an announcement
18:11 we should be adding our templates to Koha soon as well
19:30 rosa Hi anyone. is there a spanish language list for Koha?
19:46 michael ... any french here ?
19:46 hi NZ and hi FRANCE ;-)
19:46 and other too that i don't know country...
19:46 rosa hi michael
19:47 michael hi rosa
20:29 ...
20:29 ...
20:30 i'm here if anyone want to tell something great (or even) stupid in French or English ;-)
20:54 rach hi
20:54 bonjour is about the extent of my french
20:54 are you interested in Koha?
20:58 michael no pb... i wish paul was not already in bed but i'll wait or speak to him another day or night... no pb
21:45 hi...
21:45 did anyone know how u can get invited to the IRC katipo chan ??
21:58 rach hi
21:58 sorry what?
22:13 michael ...
22:13 hi rachel !
22:13 i just subscribe to the koha list and send a msg
22:14 here's the end of this message (almost in french)
22:15 the english part is :
22:15 "Koha coming soon in  bêta-test in a french library" in the subject and here in English :
22:15 ################### ###################  ###################
22:15 Hi everyone Katipo and Koha team who doesn't read or understand french,
22:15 this is for annoning that a new french library will install and test Koha
22:15 this is a public library, about 170 000 items i think and 5 000
22:15 borrowers as far as i remenber
22:15 is that possible te be added in the Koha map? i started all this stuff
22:15 of installation, migration in the next two weeks. my name is Michaël
22:15 Reichhard, i often go to chat on the IRC Channel
22:15 Here's what you can put on the Koha Map in Europe
22:15 http://koha.org/about/map/europe.html
22:15 Organisation:   Médiathèque
22:15 Version:        Koha 2.2.1
22:15 Library Type:   Public Library
22:15 Location:       Meudon, France
22:15 Contact:        Michaël Reichhard (mreichhard@gmail.com)
22:15 Description:  Bêta-test to show how Koha could be a solution for a
22:15 future migration of our old integrated library system
22:15 sorry, it's late here (00h28) and i'm not enough good to translate the
22:15 complete message. but i think Paul and i will give you more
22:15 explanations in the next few days and weeks.
22:15 Long live to open source and koha in particular ;-)
22:15 Freely,
22:15 --
22:15 Michaël
22:15 ################### ################### ###################
22:15 no joke i swear ;-)
23:18 rach hi michael, sounds good
23:18 if you're in france isn't it the middle of the night?
23:19 michael ... i hope so... i'll probably be more active in the diffrent ML
23:19 yes.. it's 3h22... always... as Albert Einstein says : "I don't sleep much. But i sleep fast"
23:19 rach :-)
23:19 michael "Je ne dors pas longtemps. Mais je dors vite"  Albert Einstein
23:20 rach paul keeps a bit more regular hours
23:20 so you wont catch him here at this time of day
23:22 that sounds like good news about your public library :-)
23:22 do you work for them, or are you a contractor?
23:26 michael yes, thanx
23:27 i work in THIS public library... 39h/semaine... anyone understand french in katipo.. if the case, i could send you the whole message in french..
23:28 rach ah no I don't think we do really I'm sorry
23:31 michael no pb... wll go to sleep soon... do you know when Paul go to Nelsonville, Ohio... did i dreamt or read this somewhere ???
23:34 i got this when i post to the list : Your mail to 'Koha' with the subject
23:34   Koha bient?t en b?ta-test dans une m?diath?que municipale du 92 /
23:34 Koha coming soon in b?ta-test in a french library
23:34 Is being held until the list moderator can review it for approval.
23:34 The reason it is being held:
23:34   Too many recipients to the message
23:37 bye... see you next time;..
23:37 at another hour maybe...
23:45 rach I think in may
01:18 anyone about? where do you find sub branches/locations
01:19 I thought we had those, or am I dreaming?
01:49 rosa we have real and virtual branches
10:22 sylvain hi all

← Previous day | Today | Next day → | Search | Index

koha1