sql speed

[jda]

>Hi Jon,
>
>>I have put together a method of joining cursors by storing recIDs 
>>of individual searches in arrays, and then either ANDing or ORing 
>>the arrays (in RB). I then do a final search using the recIDs that 
>>are in the final array. As a result, there can be thousands of 
>>numbers. I find that the search is actually pretty slow, no doubt 
>>due to the sql overhead. For the search below, for about 2800 
>>recIDs, it takes a about 7 seconds. I wonder if there is something 
>>new in V2 that would speed this up (I still need a cursor at the 
>>end of the operation). Thanks.
>
>I suggest looking at VBitSet and VSet in the V4RB Reference PDF. You 
>can use the Union method to merge your sets and it will be much 
>faster. However, I don't see an obvious way to convert it into a 
>VCursor as a last step. That should be possible I think. Maybe you 
>can Feature Request something. :^)

Hi Frank,

Yes, good idea. I thought of VSets, but I do need a cursor at the end...

>
>Maybe a ToCursor Method with a variable number of parameters 
>representing fields in the Table.
>
>Vset.ToCursor(field1, field2, ... fieldx) As VCursor
>
>fieldx could be a text string with the field name.
>
>>select 
>>id,hit,uniqueID,authors,thedate,title,editors,authorsSortMethod 
>>from thereferences where RecID in 
>>(30,33,34,38,39,40,43,56,57,61,62,63,64,65,66,70,71,72,73,86,87,92,114,115,142,157,163,167,169,171,172,174,175,176,178,181,216,219,225,226,235,236,239,240,241,263,264,266,270,271,272,276,277,279,297,298,304,306,307,308,327,328,331,332,333,334,335,336,337,338,341,342,343,344,345,346,348,349,350,351,358,360,362,376,377,396,398,400,405,406,407,414,419,420,434,436,437,438,441,443,456,457,458,473,475,486,488,493,499,502,509,511,522,524,529,530,538,544,546,547,551,555,556,563,567,569,574,575,581,582,583,584,585,591,605,616,618,633,634,635,636,637,639,650,671,695,697,698,699,700,701,702,703,704,705,707,709,710,711,719,720,721,722,735,742,743,744,745,755,756,757,774,775,776,781,784,810,813,814,815,822,833,849,856,864,865,866,867,868,869,876,888,889,890,891,892,893,894,895,912,914,918,922,923,935,941,960,961,963,964,966,967,968,969,970,996,997,1022,1023,1024,1025,1026,1036,1037,1038,1043,1044,1045,1046,1047,1048,1049,1051,1052,1053,1054,1058,1059,1060,1061,1062,1063,1064,1065,1069,!
 10!
>
>What's that ! character doing in there? It seems to end every 
>paragraph of numbers.


Must be an email client thing...it's not there in the RB debugger. I 
think Eudora puts it in to break up long lines without a return,



Thanks,

Jon