Reinhard Doelz (doelz at comp.bioz.unibas.ch) wrote:
: : Does anyone know of pre-existing scripts that could be run as a cron job
: Two considerations have to be kept in mind:
: 1) The update files become very large (100MB) and a considerable scratch
: space is required if you download all at once (.Z is 2/3rd of the original
: flat file, and the flat file must co-exist with the gcg file spacewise).
: Including a save copy of GCG formatted data the currently needed space,
: therefore, is about 350 MByte by now. Think about the growth and calcu-
: late for 2 Gigabyte at the end of next year.
: 2) Incremental updating cannot simply append files because of duplications
: and deletions. To my knowledge, there is no current GENBANK-compliant
: procedure which will allow processing of the incremental updates and its
: quality validation.
Your point about space considerations is well taken; however, I am not
aware of any other way of performing (for example) fasta searching in
real time on an up-to-date db. Since hard disk costs have dropped to 20
cents a meg and less, storing db locally is likely to be financially
viable for some time to come, even given the projected sizes of the major
db's. Presumably it's in the works at GCG to run some kind of central db
that folk can connect to in real time, some time in the future, much like
BLAST searching at NCBI can be done.