>>>>Hi All. I dusted off an old copy of PHDbase :) and tried to make it work with VFP9 but it always gives me an error after creating the index. I know its old and all that but doesn't the UT use this to perform searches? Anyone know if its even still compatible with VFP9? Just curious.
>>>
>>>What is the version you have?
>>
>>3.61. Whats the latest and how would one get it now?
>
>3.61 is the last version Korenthal ever made, AFAIK. My test has shown that phDbase crashes VFP if you have too many records, and I have also found that it does not find all the matching records, even after a reindexing. So for my own personal needs, I created a few functions to get the features I needed.


Hi Tore and Michel. I also wrote some routines a few years ago that does this. I guess we all have in order to solve this problem!

My routines create an index in the source table which calls a function that updates a master words table and a link back table containing the links back to the source table. The use of an index on the source table allows the word master tables to automatically update as new records are added or removed from the source table or as the contents of the fields being indexed changes. This is usefull.

It records word frequency and statistics as well. Most important is that it updates as the source table changes. The downside is that the words master file and link back tables plus their indexes were as big as the source table which effectively doubled the space requirements :( It is also an entirely VFP solution except for a small dll that a collegue wrote in assembler to parse all the words out of the source fields being indexed. This was for speed so that updates could be done fast in real time as records were added or changed.

Michel, I can send you this code for you to see whether its useful to you/UT. (Don't laugh at it though - I know that it's not very professonal - <g>).