xbrowser footer

[wartiaga]

Hello,

Why this:

Lista2:aCols[1]:cFooter := { || Ltrim( Str( Lista2:KeyNo() ) ) + " / " + LTrim( Str( Lista2:KeyCount() ) ) }
Lista2:Refresh()

Show me {||...} in footer?

Thanks in advance.

[cnavarro]

Hello,

Why this:

Lista2:aCols[1]:cFooter := { || Ltrim( Str( Lista2:KeyNo() ) ) + " / " + LTrim( Str( Lista2:KeyCount() ) ) }
Lista2:Refresh()

Show me {||...} in footer?

Thanks in advance.

Use this

Code: Select all

Lista2:aCols[1]:bFooter   := { || Ltrim( Str( Lista2:KeyNo() ) ) + " / " + LTrim( Str( Lista2:KeyCount() ) ) }
 

[wartiaga]

Hello,

Why this:

Lista2:aCols[1]:cFooter := { || Ltrim( Str( Lista2:KeyNo() ) ) + " / " + LTrim( Str( Lista2:KeyCount() ) ) }
Lista2:Refresh()

Show me {||...} in footer?

Thanks in advance.

Use this

Code: Select all

Lista2:aCols[1]:bFooter   := { || Ltrim( Str( Lista2:KeyNo() ) ) + " / " + LTrim( Str( Lista2:KeyCount() ) ) }
 

Thank you, works great, but Lista2:KeyCount() is counting all records and i use a scope in lista2, how i can display the real count (not count deleted)? Thank you again.

[cnavarro]

Try with

Code: Select all

oBrw:aCols[ 1 ]:nFooterType := AGGR_COUNT
oBrw:aCols[ 1 ]:bSumCondtion := { |v,o| ( oBrw:cAlias ) -> ( !Deleted() ) }
 

[nageswaragunupudi]

oBrw:KeyCount() is the same as ( oBrw:cAlias )->( OrdKeyCount() )

OrdKeyCount() is a (x)Harbour function.

By default OrdKeyCount() respects Scopes and Filters but includes deleted records also. This is what we all should know.

If we want OrdKeyCount() not to include deleted records, we should set filter to "SET FILTER TO !DELETED()"

If we are particular that oBrw:nLen or oBrw:KeyCount() should not include deleted records then

Code: Select all

USE MYTABLE ....
SET FILTER TO !DELETED()
GO TOP
// over and above this, we can use Scopes also.

// now build xbrowse
 

If we want to set some filter like "AGE > 40", then set filter like this:

Code: Select all

SET FILTER TO AGE > 47 .AND. !DELETED()
 

OrdKeyNo() also is effected in the same way.

Note:
1. Using oBrw:nLen is faster than using oBrw:KeyCount().
2. These are issues relating to using the (x)Harbour function OrdKeyCount() but not XBrowse.

[wartiaga]

Try with

Code: Select all

oBrw:aCols[ 1 ]:nFooterType := AGGR_COUNT
oBrw:aCols[ 1 ]:bSumCondtion := { |v,o| ( oBrw:cAlias ) -> ( !Deleted() ) }
 

Thank you!

[wartiaga]

oBrw:KeyCount() is the same as ( oBrw:cAlias )->( OrdKeyCount() )

OrdKeyCount() is a (x)Harbour function.

By default OrdKeyCount() respects Scopes and Filters but includes deleted records also. This is what we all should know.

If we want OrdKeyCount() not to include deleted records, we should set filter to "SET FILTER TO !DELETED()"

If we are particular that oBrw:nLen or oBrw:KeyCount() should not include deleted records then

Code: Select all

USE MYTABLE ....
SET FILTER TO !DELETED()
GO TOP
// over and above this, we can use Scopes also.

// now build xbrowse
 

If we want to set some filter like "AGE > 40", then set filter like this:

Code: Select all

SET FILTER TO AGE > 47 .AND. !DELETED()
 

OrdKeyNo() also is effected in the same way.

Note:
1. Using oBrw:nLen is faster than using oBrw:KeyCount().
2. These are issues relating to using the (x)Harbour function OrdKeyCount() but not XBrowse.

Thank you Mr. Nages, I use ! Deleted() in index condition to solve this issue.

[nageswaragunupudi]

Thank you Mr. Nages, I use ! Deleted() in index condition to solve this issue.

Please do NOT use FOR conditions in Indexes.

The purpose of indexes is also to speed up queries (FILTERS). The RDD does not consider indexes with FOR conditions for optimization. That way we forgo some benefits of having indexes.

As I suggested please use FILTER TO !DELETED().

There is not much performance impact if the percentage of deleted records is less.

Create an additional index tag on every large dbf as

Code: Select all

INDEX ON DELETED() TAG DELETED
 

Then the filter "SET FILTER TO !DELETED()" also is fully optimized.

Better we familiarize ourselves with the way RDD optimizes filters.

[Enrico Maria Giordano]

I can't see any optimization in the following sample:

Code: Select all

#include "Fivewin.ch"


REQUEST DBFCDX


FUNCTION MAIN()

    LOCAL nSec

    LOCAL n, i

    RDDSETDEFAULT( "DBFCDX" )

    SET OPTIMIZE ON

    DBCREATE( "MYTEST", { { "TEST", "C", 100, 0 } } )

    USE MYTEST

    FOR i = 1 TO 300000
        APPEND BLANK

        REPLACE FIELD -> test WITH "TEST" + LTRIM( STR( i ) )

        IF i % 2 = 0
            DELETE
        ENDIF
    NEXT

    INDEX ON FIELD -> test TAG TEST TO MYTEST FOR !DELETED()

    nSec = SECONDS()

    n = 0

    WHILE !EOF()
        n++
        SKIP
    ENDDO

    ? n, SECONDS() - nSec

    CLOSE

    FERASE( "MYTEST.CDX" )

    USE MYTEST

    INDEX ON DELETED() TAG DELETED TO MYTEST
    INDEX ON FIELD -> test TAG TEST TO MYTEST

    SET FILTER TO !DELETED()

    nSec = SECONDS()

    n = 0

    WHILE !EOF()
        n++
        SKIP
    ENDDO

    ? n, SECONDS() - nSec

    CLOSE

    FERASE( "MYTEST.DBF" )
    FERASE( "MYTEST.CDX" )

    RETURN NIL

Results: 0.07 and 1.06.

Can you change my sample to show the optimization, please?

EMG

[Enrico Maria Giordano]

Rao, have you had the time to test my sample?

EMG