Cohousing-L Mailing List: Re: Re: coop/condo financing

[apguirard (apguirard]

We could create a search entry screen to find records by description _alone_, 
constructing the query ourselves, but even so it's not an easy thing.  Consider 
that if the search string is: "*ABC,*EEQ,*PDQ*", we'd want to match records 
with any of the following:

   1: ABC,EEQ,PDQ

   1: HAHA,ABC,NEVERMIND,LA,DE,DA,
   2: EEQ,12.003,PDQ,WHAT?

   1: SOMETHING,SOMETHING,SOMETHING,
   2: HAHA,ABC,NEVERMIND,LA,DE,DA,
   3: 23,EEQ,12.003,DON'T BUG ME,
   4: PDQ

Since each line is a separate UDA, we'd have to check whether each line matched 
every possible subexpression.  Unless the information is
all together is a single field, the matching would be hard to do.  We'd have to 
break the search string up at stars and commas and check for each word 
individually, thus:

uda.{name="PART_DESC*" and value="*ABC,*"} and uda.{name="PART_DESC*" and 
value="*EEQ,*"} and uda.{name="PART_DESC*" and value="*PDQ,*"}

We'd then have to take each match, read the record in, paste together the long 
part description, and test against the original search string.  Note that 
though there may be only one record that has the requested substrings in the 
order the user specifies, the search above will find them even if they're out 
of order.  Depending on the query, that may be a lot of records.  It could take 
a while.

What is the object of this multi-line search, incidentally?  Can you give me 
some examples where it would be used?  Isn't the information stored elsewhere 
in the record where it could be got to on a field by field basis?  How would it 
be if we just let the DESC field hold more characters?  Then they could search 
on that and it would at least improve the situation.