Excel 97

R

Rodney

During a VLookup,
(A 1400 name list, looks for matches in a 150,000 array)

The seek time takes just under 12 minutes,
there is no "hourglass" indicating the program is computing
The "calculating time" stays on 0% until the operation
is about over then quickly spins through the percentage table
to finish.
Is this roughly a correct scenario?

I have a rather flaccid 386 celeron with ASUS MB
and 300MB ram.

Thx
 
M

Max

FWIW, yes.

I'm also in xl97 and have come across the same Excel
behaviour you posted, whenever there's a terrific amount
of calcs involved (with a PIII-500 with 128Mb RAM)

Most important is that correct, calculated results
are returned at the end of it.
 
R

Rodney

Thanks Max,
we use MSWorks for the majority of our work
and the simple program moves along at light speed
so we get a bit spoilt :)



--
(e-mail address removed)
(Remove gum to reply)


| FWIW, yes.
|
| I'm also in xl97 and have come across the same Excel
| behaviour you posted, whenever there's a terrific amount
| of calcs involved (with a PIII-500 with 128Mb RAM)
|
| Most important is that correct, calculated results
| are returned at the end of it.
|
| --
| Rgds
| Max
| xl 97
| ---
| Please respond, in newsgroup
| xdemechanik <at>yahoo<dot>com
| ----
|
| "Rodney" wrote:
| >During a VLookup,
| >(A 1400 name list, looks for matches in a 150,000 array)
| >
| >The seek time takes just under 12 minutes,
| >there is no "hourglass" indicating the program is
| computing
| >The "calculating time" stays on 0% until the operation
| >is about over then quickly spins through the percentage
| table
| >to finish.
| >Is this roughly a correct scenario?
| >
| >I have a rather flaccid 386 celeron with ASUS MB
| >and 300MB ram.
| >
| >Thx
| >
| >
| >--
| >[email protected]
| >(Remove gum to reply
 
R

RagDyeR

Vlookups are notoriously slow, especially if you're error trapping using
double and/or triple Vlookup formulas.

I have a few old, very large data base type WSs which *had* exhibited calc
times similar to what you are seeing.

I've since revised them, based on information shared within these NGs, and
have been able to reduce the calc times, in some cases, to a better then 75%
improvement.

The main improvement was realized when changing the formulas from Vlookup to
Index and Match, and eliminating the double Vlookups for error trapping.

I would advise you to experiment with revising your formula approach, and
see if a function change can help you.

If you wish, post some of your largest (long, complex) Vlookup formulas
here, and lets see if we can help with suggesting shorter calc time
alternatives.
--

HTH,

RD
==============================================
Please keep all correspondence within the Group, so all may benefit!
==============================================

Thanks Max,
we use MSWorks for the majority of our work
and the simple program moves along at light speed
so we get a bit spoilt :)



--
(e-mail address removed)
(Remove gum to reply)


| FWIW, yes.
|
| I'm also in xl97 and have come across the same Excel
| behaviour you posted, whenever there's a terrific amount
| of calcs involved (with a PIII-500 with 128Mb RAM)
|
| Most important is that correct, calculated results
| are returned at the end of it.
|
| --
| Rgds
| Max
| xl 97
| ---
| Please respond, in newsgroup
| xdemechanik <at>yahoo<dot>com
| ----
|
| "Rodney" wrote:
| >During a VLookup,
| >(A 1400 name list, looks for matches in a 150,000 array)
| >
| >The seek time takes just under 12 minutes,
| >there is no "hourglass" indicating the program is
| computing
| >The "calculating time" stays on 0% until the operation
| >is about over then quickly spins through the percentage
| table
| >to finish.
| >Is this roughly a correct scenario?
| >
| >I have a rather flaccid 386 celeron with ASUS MB
| >and 300MB ram.
| >
| >Thx
| >
| >
| >--
| >[email protected]
| >(Remove gum to reply
 
R

Rodney

Thanks Don,
I have to keep it relatively simple,
I have a partner in Bangkok who would baulk
at using Macro's. (That's if it would work in our case)

RD, thanks for your reply,
I have "worked around" by using 5 templates
each using 30,000 or so in the array
and using this formula.........
=IF(ISERROR(VLOOKUP(B47063,$B$2:$C$47052,2,FALSE)),"x",(VLOOKUP(B47063,$B$2:$C$47052,2,FALSE)))

The seek time is about 30 seconds for each,
however when I had 5 rows (arrays) the resultant time was 12 minutes

Regards





--
(e-mail address removed)
(Remove gum to reply)


| Vlookups are notoriously slow, especially if you're error trapping using
| double and/or triple Vlookup formulas.
|
| I have a few old, very large data base type WSs which *had* exhibited calc
| times similar to what you are seeing.
|
| I've since revised them, based on information shared within these NGs, and
| have been able to reduce the calc times, in some cases, to a better then 75%
| improvement.
|
| The main improvement was realized when changing the formulas from Vlookup to
| Index and Match, and eliminating the double Vlookups for error trapping.
|
| I would advise you to experiment with revising your formula approach, and
| see if a function change can help you.
|
| If you wish, post some of your largest (long, complex) Vlookup formulas
| here, and lets see if we can help with suggesting shorter calc time
| alternatives.
| --
|
| HTH,
|
| RD
| ==============================================
| Please keep all correspondence within the Group, so all may benefit!
| ==============================================
|
| | Thanks Max,
| we use MSWorks for the majority of our work
| and the simple program moves along at light speed
| so we get a bit spoilt :)
|
|
|
| --
| (e-mail address removed)
| (Remove gum to reply)
|
|
| | | FWIW, yes.
| |
| | I'm also in xl97 and have come across the same Excel
| | behaviour you posted, whenever there's a terrific amount
| | of calcs involved (with a PIII-500 with 128Mb RAM)
| |
| | Most important is that correct, calculated results
| | are returned at the end of it.
| |
| | --
| | Rgds
| | Max
| | xl 97
| | ---
| | Please respond, in newsgroup
| | xdemechanik <at>yahoo<dot>com
| | ----
| |
| | "Rodney" wrote:
| | >During a VLookup,
| | >(A 1400 name list, looks for matches in a 150,000 array)
| | >
| | >The seek time takes just under 12 minutes,
| | >there is no "hourglass" indicating the program is
| | computing
| | >The "calculating time" stays on 0% until the operation
| | >is about over then quickly spins through the percentage
| | table
| | >to finish.
| | >Is this roughly a correct scenario?
| | >
| | >I have a rather flaccid 386 celeron with ASUS MB
| | >and 300MB ram.
| | >
| | >Thx
| | >
| | >
| | >--
| | >[email protected]
| | >(Remove gum to reply
|
|
|
 
R

RagDyeR

The formula you posted is a classic example of a double Vlookup which bloats
the calc times of *large* WSs.

I would appreciate you trying the following formula, and posting back with
the calc times, as compared to your original one.
There has to be a definite improvement.

=IF(ISNA(MATCH(B47063,B2:B47052,0)),"x",INDEX(C2:C47052,MATCH(B47063,B2:B470
52,0)))

A similar formula that I revised in my large WS saved almost 50% in calc
time, where a triple Vlookup, as I said before, saved almost 75%.
--

HTH,

RD
==============================================
Please keep all correspondence within the Group, so all may benefit!
==============================================


Thanks Don,
I have to keep it relatively simple,
I have a partner in Bangkok who would baulk
at using Macro's. (That's if it would work in our case)

RD, thanks for your reply,
I have "worked around" by using 5 templates
each using 30,000 or so in the array
and using this formula.........
=IF(ISERROR(VLOOKUP(B47063,$B$2:$C$47052,2,FALSE)),"x",(VLOOKUP(B47063,$B$2:
$C$47052,2,FALSE)))

The seek time is about 30 seconds for each,
however when I had 5 rows (arrays) the resultant time was 12 minutes

Regards





--
(e-mail address removed)
(Remove gum to reply)


| Vlookups are notoriously slow, especially if you're error trapping using
| double and/or triple Vlookup formulas.
|
| I have a few old, very large data base type WSs which *had* exhibited calc
| times similar to what you are seeing.
|
| I've since revised them, based on information shared within these NGs, and
| have been able to reduce the calc times, in some cases, to a better then
75%
| improvement.
|
| The main improvement was realized when changing the formulas from Vlookup
to
| Index and Match, and eliminating the double Vlookups for error trapping.
|
| I would advise you to experiment with revising your formula approach, and
| see if a function change can help you.
|
| If you wish, post some of your largest (long, complex) Vlookup formulas
| here, and lets see if we can help with suggesting shorter calc time
| alternatives.
| --
|
| HTH,
|
| RD
| ==============================================
| Please keep all correspondence within the Group, so all may benefit!
| ==============================================
|
| | Thanks Max,
| we use MSWorks for the majority of our work
| and the simple program moves along at light speed
| so we get a bit spoilt :)
|
|
|
| --
| (e-mail address removed)
| (Remove gum to reply)
|
|
| | | FWIW, yes.
| |
| | I'm also in xl97 and have come across the same Excel
| | behaviour you posted, whenever there's a terrific amount
| | of calcs involved (with a PIII-500 with 128Mb RAM)
| |
| | Most important is that correct, calculated results
| | are returned at the end of it.
| |
| | --
| | Rgds
| | Max
| | xl 97
| | ---
| | Please respond, in newsgroup
| | xdemechanik <at>yahoo<dot>com
| | ----
| |
| | "Rodney" wrote:
| | >During a VLookup,
| | >(A 1400 name list, looks for matches in a 150,000 array)
| | >
| | >The seek time takes just under 12 minutes,
| | >there is no "hourglass" indicating the program is
| | computing
| | >The "calculating time" stays on 0% until the operation
| | >is about over then quickly spins through the percentage
| | table
| | >to finish.
| | >Is this roughly a correct scenario?
| | >
| | >I have a rather flaccid 386 celeron with ASUS MB
| | >and 300MB ram.
| | >
| | >Thx
| | >
| | >
| | >--
| | >[email protected]
| | >(Remove gum to reply
|
|
|
 
R

RagDyeR

Forgot the absolutes for the list ranges:

=IF(ISNA(MATCH(B47063,$B$2:$B$47052,0)),"x",INDEX($C$2:$C$47052,MATCH(B47063
,$B$2:$B$47052,0)))

--

Regards,

RD
--------------------------------------------------------------------
Please keep all correspondence within the Group, so all may benefit !
--------------------------------------------------------------------

The formula you posted is a classic example of a double Vlookup which bloats
the calc times of *large* WSs.

I would appreciate you trying the following formula, and posting back with
the calc times, as compared to your original one.
There has to be a definite improvement.

=IF(ISNA(MATCH(B47063,B2:B47052,0)),"x",INDEX(C2:C47052,MATCH(B47063,B2:B470
52,0)))

A similar formula that I revised in my large WS saved almost 50% in calc
time, where a triple Vlookup, as I said before, saved almost 75%.
--

HTH,

RD
==============================================
Please keep all correspondence within the Group, so all may benefit!
==============================================


Thanks Don,
I have to keep it relatively simple,
I have a partner in Bangkok who would baulk
at using Macro's. (That's if it would work in our case)

RD, thanks for your reply,
I have "worked around" by using 5 templates
each using 30,000 or so in the array
and using this formula.........
=IF(ISERROR(VLOOKUP(B47063,$B$2:$C$47052,2,FALSE)),"x",(VLOOKUP(B47063,$B$2:
$C$47052,2,FALSE)))

The seek time is about 30 seconds for each,
however when I had 5 rows (arrays) the resultant time was 12 minutes

Regards





--
(e-mail address removed)
(Remove gum to reply)


| Vlookups are notoriously slow, especially if you're error trapping using
| double and/or triple Vlookup formulas.
|
| I have a few old, very large data base type WSs which *had* exhibited calc
| times similar to what you are seeing.
|
| I've since revised them, based on information shared within these NGs, and
| have been able to reduce the calc times, in some cases, to a better then
75%
| improvement.
|
| The main improvement was realized when changing the formulas from Vlookup
to
| Index and Match, and eliminating the double Vlookups for error trapping.
|
| I would advise you to experiment with revising your formula approach, and
| see if a function change can help you.
|
| If you wish, post some of your largest (long, complex) Vlookup formulas
| here, and lets see if we can help with suggesting shorter calc time
| alternatives.
| --
|
| HTH,
|
| RD
| ==============================================
| Please keep all correspondence within the Group, so all may benefit!
| ==============================================
|
| | Thanks Max,
| we use MSWorks for the majority of our work
| and the simple program moves along at light speed
| so we get a bit spoilt :)
|
|
|
| --
| (e-mail address removed)
| (Remove gum to reply)
|
|
| | | FWIW, yes.
| |
| | I'm also in xl97 and have come across the same Excel
| | behaviour you posted, whenever there's a terrific amount
| | of calcs involved (with a PIII-500 with 128Mb RAM)
| |
| | Most important is that correct, calculated results
| | are returned at the end of it.
| |
| | --
| | Rgds
| | Max
| | xl 97
| | ---
| | Please respond, in newsgroup
| | xdemechanik <at>yahoo<dot>com
| | ----
| |
| | "Rodney" wrote:
| | >During a VLookup,
| | >(A 1400 name list, looks for matches in a 150,000 array)
| | >
| | >The seek time takes just under 12 minutes,
| | >there is no "hourglass" indicating the program is
| | computing
| | >The "calculating time" stays on 0% until the operation
| | >is about over then quickly spins through the percentage
| | table
| | >to finish.
| | >Is this roughly a correct scenario?
| | >
| | >I have a rather flaccid 386 celeron with ASUS MB
| | >and 300MB ram.
| | >
| | >Thx
| | >
| | >
| | >--
| | >[email protected]
| | >(Remove gum to reply
|
|
|
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top