Does the quantity of text or the number of records or searches justify all the additional code and processes?
VFP is wickedly fast at text manipulation and searching. searching for "Rabbit" in the full text (available from Project Gutenberg) returns 101 records out of 101 in 0.03 seconds on an old and small machine.
CREATE DATABASE textsearch CREATE TABLE textsearch (iKey int autoinc, mText M, tUpdated T DEFAULT DATETIME()) FOR I=1 TO 101 INSERT INTO textsearch (mText) VALUES (textfile) NEXT SELECT * FROM textsearch WHERE "Rabbit" $ mText * returned 101 records in 0.03 seconds from a 15 Mb fpt
FOR I=1 TO 1000 INSERT INTO textsearch (mText) VALUES (textfile) NEXT SELECT * FROM textsearch WHERE "Rabbit" $ mText * returned 1101 records in 0.33 seconds from a 150 Mb FPT * quit VFP and restart for cache clearing: 0.63 seconds
On Wed, Aug 1, 2018 at 11:05 AM, Alan Bourke alanpbourke@fastmail.fm wrote:
I was considering a simple MSSQL table with an nvharchar(max) field to
hold the VFP memo contents, and full-text indexing on it. Then it is just a question of querying that from VFP, and refreshing the records in it from VFP on some sort of timed basis.
-- Alan Bourke alanpbourke (at) fastmail (dot) fm
[excessive quoting removed by server]