LOOKUP() is in deed not Rushmore optimizable, but if you include all four parameters, that really does not matter. Rushmore is the name for VFP's algorithm to try to select the best use of the existing indexes, but if you specify which index to use, you take over this job yourself. Since also SEEK and SEEK() requires that you have specified an index, Rushmore is per definition inactive when you use these functions/commands. At least that is how I interpret the docs.


Hi Tore, VFP docs is a start point, nothing more.

LOOKUP use a special optimization:
- if the fourth parameter exist, it have precedency,
for the third parameter it must to be a valid fieldname
- if for the third parameter exist a tag, it use the tag
- otherwise it do a linear scan like LOCATE

CREATE CURSOR myCursor (F1 C(1),F2 I,F3 Y )
INSERT INTO myCursor VALUES ('A',3,$4)
INSERT INTO myCursor VALUES ('B',2,$5)

INDEX ON F1 TAG tagF1
INDEX ON F2 TAG tagF2
INDEX ON F1+BINTOC(F2) TAG tagF1F2

?LOOKUP(F1,'A',F1)                       && THIS USE A SEEK IN THE TAG tagF1
?LOOKUP(F1,3,F2)                         && THIS USE A SEEK IN THE TAG tagF2
?LOOKUP(F1,$4,F3)                        && THIS USE A SCAN LOOKUP
? LOOKUP(F1,'A'+BINTOC(3),F3,'tagF1F2')   && THIS USE A SEEK IN THE TAG tagF1F2
? LOOKUP(F1,'A'+BINTOC(3),F4,'tagF1F2')   && error