vortex select syntax

[resume.robot]

This string works

"select 'http://' + Url Url, Title, Body, length(Body) Size, id, Visited from html where Title\Meta\Body likep $query and (Title\Meta\Body like '(resume,resum\351,curriculum vitae,cv,bio,profile)')">

This one doesn't

"select 'http://' + Url Url, Title, Body, length(Body) Size, id, Visited from html where Title\Meta\Body likep $query and (Title\Meta\Body like '((resume education),resum\351,curriculum vitae,cv,bio,profile)')">


The difference is nested parentheses (resume education)

The entry I want to select must contain BOTH the words resume and education, or ONE of the other terms.

What is the correct way to write this?

Thanks

Mike Clark

[mark]

Drop the extra parens. A term with space in a paren list is inherently a phrase which must match entirely. e.g. curriculum vitae is also a 2 term phrase.

[resume.robot]

Thank you

But I don't want the phrase, I want both terms occurring somewhere in the body, not necessarily near each other.

Is it possible to select (a AND b) OR (c AND d) OR e OR f?

[mark]

Sounds like you need some boolean logic:
... likep $query and (Title\Meta\Body like '(resume education,resum\351,cv,bio,profile)' or Title\Meta\Body like 'curriculum vitae')

There may be better ways to do this overall depending on what you're really doing. If that is a constant list of terms it would be better to perform one pass over the database to flag records matching those sets so you only need to do one simple comparison at search time.

[resume.robot]

Thank you. Sorry to be dense, please let me explain exactly what we need. We want to display resumes and eliminate job postings.

First, we need to delete from the database, or exclude from being displayed, any record that contains ONE of the following terms


eoe
equal-opportunity-employer
resume-writing
resume-guide
resume-service
resume-distribution
resume-posting
resume-scanning
resume-utility
peoplesearch
people-search
jobbankusa
job-bank-usa
sample-resume
submit-resume
email-resume
fax-resume
send-resume
your-resume
post-resume
jobs.offered
jobs.agency
employment-agency
this-position
positions-vacant
following-positions
vacancies
please-apply
career-opportunities
resume-format
position-available
positions-available
job-location
job-summary
how-to-apply
position-open
positions-open
forward-resume
searching-for-people
we-are-seeking
required-skills
skills-required

Then we need to display in searches all records which contain the following terms

(resume,resum\351,curriculum vitae,cv,curriculum vitae)

Here is one problem: The word resume has multiple meanings. We want to include results which contain (resume AND education) OR (resume AND experience). This is not a phrase, just 2 words on the same page.

We are a full texis licensee. However, the machine on which this database is currently running uses commercial webinator. Ultimately it will be moved to the 64 bit sun machine with full texis. If that will make the database easier to manage for this purpose, we can move it sooner. What is your recommendation?

Thanks

Mike Clark

[mark]

If there are records you never want you should delete them:
gw -st "delete from html where Title\Meta\Body like '(skills required,required skills,etc...)'"

If I follow correctly you want records matching
Title\Meta\Body like 'cv resum\351 curriculum-vitae @0' or
Title\Meta\Body like 'resume (education,experience)' or
Or to delete the bad ones:
delete from html where not (
-- the 2 likes from above
)

If you want to keep the records, rather than deleting them, you should flag them once so the search is simpler (just check the flag). If you have to stick with the gw walker you can use the "New" field which gw doesn't use. Otherwise you can use the scripted crawler ftp://ftp.thunderstone.com/pub/dowalk_beta with your full texis and expand the schema to add columns as needed.

[doran]

Note to other Webinator admin's -- this user is running full Texis, as appropriate for such an application, although using gw for some data gathering....

[mark]

It's still fundamentally a webinator app for technical purposes. This was as good a place as any for it. Though the license note was worth making.