splitting a field

Frederic Achard Frederic.Achard at infobiogen.fr
Tue Oct 24 14:00:40 EST 1995


I am trying to index a new data bank for our srs
server and I am not sure about the proper way to
do so. My bank is a bank of links and an entry
looks like

" LNK   vgl: 1        gdb-gbk  link GDB_gene and Genbank_sequence
" GO1   gdb: 118727   NPPA natriuretic peptide precursor A
" GO2   gbk: M54951   HUMANFZ2 Human atrial natriuretic factor gene, 3'
" COM   /score=0.56  /author=genXref v0.9  /date=24-Oct-1995
" //

LNK is the link field
GO1 and GO2 are genomic objects 1 and 2
COM are comments on that field
And the identifier of a data bank object is
made of a mnemonic (-3 letters, eg. gbk is
Genbank,) a ':',  a space and an unique id

The main index should be on the number following
"LNK   vgl:"

I want to have the following stuffs indexed:

   - the database-ids which are taken from
      LNK, GO1 and GO2 fields ('vgl: 1',
      'gdb: 118727' and 'gbk: M54951')

   - keywords taken from GO1 and GO2 (but excluding
     the previous database-id)

In genbank, I saw that two tokens are taken from the
LOCUS field (locus name and date) but couldn't really
figure out how this works.
Does it work because \find="" on the #libformat of date
or (and) because a function  (f=@F_SAVEDATEGB) is called
to transmogrify the date in the locus parser.

By the way, how do I create my function. Just add the code
in srsbuild.c and compile, section and update everything ???

I'd REALLY appreciate any help. A pointer on anything
similar will do.

I thank you in advance.


PS: we have a srs for development on our site.
You can check what I have done so far at
in virgil.

PS 2: sorry but the <reply to> of my previous post was wrong.
Hope it did not bother any of you. Does it explain why I got no
response ;-)

PS3: I forgot the subject on the previous one. So I repost it. 
Not sure if this is good Internet usage .

More information about the Bio-srs mailing list

Send comments to us at biosci-help [At] net.bio.net