>>>Hello,
>>>
>>>I have a table "student" that tracks which day a particular student takes a lesson, what time, which room, and with which instructor. In other words the pertinent fields are stu_day, stu_room, stu_bgntime and stu_endtime.
>>>
>>>I have an index that is called stu_day which is str(stu_day)+stu_room+str(stu_bgntime).
>>>
>>>Lets say I have a student scheduled for day 1, room 1 , and from 1:00pm to 2:00pm.
>>>
>>>My problem is ... that if I set order to stu_day and do a seek on day 1, room 1 and bgntime of 1:30 I would like this particular student to be selected, However the index is set on stu_bgntime so this student will not be selected. If I "Set Near on" it will find the next record. Is there a way to do a find <= on an index in VFP. Or is the best way to do this to "Set Near On" and then if !found() do a seek -1 and then do some checking to make sure the time is "Between" stu_bgntime and stu_endtime and the day and room are correct?
>>>
>>>Sorry if this is a stupid question.
>>
>>I have a similar situation for which I must seek on 3 values, 2 of which are known and the third falling within a range. Basically, you can't use SEEK for a situation like this, so I use LOCATE FOR, and use the 3 different parameters. If you have indexes on all of the values you LOCATE on, it will work pretty quickly, but not nearly as quickly as it would be with a seek. On a table on the network with about 400,000 records, this locate line takes about 4 seconds. I would be interested to hear if there is a better solution.
>>
>>quote from NFL commercial:
>>"There is no such thing as a stupid question. Just stupid people asking questions."
>>
>>Really though, the only stupid question is one asked before even attempting to think through the answer one's self. Your's is not a stupid question.
>>
>>Erik
>
>I don't see why seek cannot be used:
>Select table1
>Set order to key1
>Set near on
>Seek somevalue && includes beginning value of the range
>DO WHILE rangevalue<=endrangevalue
> **** count records, or check for the first one, or populate cursor with records, etc.
> Select table1
> SKIP
>ENDDO

Sounds like a better idea than mine... I think I'll be changing my code. Thanks Edward...