error 1073741824

[bparksb]

I am getting an ERROR 1073741824

Code: Select all

LS DRINK C SPIR_TAX LOG(SPIR_TAX) 1/SPIR_TAX SPIR_TAX^2

works fine but

Code: Select all

LS DRINK C 1/SPIR_TAX

gets a dialogue box with error 1073741824

Data set at
http://econ413.wustl.edu/eviews_1073741824.wf1

Eviews 8 standard edition Nov 19 build

[EViews Gareth]

I cannot replicate this. Could you state exactly what you did?

[startz]

I replicated Bob's bug by opening the workfile and typing in the command. I did get another, apparently random, number in the alert box. This was on my home machine. On my work machine the code runs fine. (Both are the current build of EViews.)

[EViews Gareth]

Could you do the supportinfo command on both machines, and paste results here (or email)?

[bparksb]

Here is my support info

=======
PRODUCT
=======
Version: EViews 8 32-bit
Build date: Nov 19 2013
Serial Number:

=========
LICENSING
=========
Standalone : C:\Documents and Settings\All Users\Application Data\IHS EViews\EViews\8.00.000\license.txt

=====
PATHS
=====
Installation path: C:\Program Files\EViews 8
Working path: G:\E413\ADATA
INI path: C:\Documents and Settings\xxxxxxx\Application Data\Quantitative Micro Software\EViews
DB INI path: C:\Documents and Settings\xxxxxxx\Application Data\Quantitative Micro Software\EViews
Help path: C:\Program Files\EViews 8\Help Files

===========
ENVIRONMENT
===========
OS: Windows XP Service Pack 3 32Bit
Total Address Space: 1453476k
VMEM Block Space: 1228324k
VMEM Space In Use: 24892k

EViews Address Space: 1206888k
EV Add Space In Use: 3456k
Cores in use: 2

[bparksb]

I just did a reboot and opened Eviews as the first program I obtained identical results - error number is the same 1073741824
Support info same.

Code: Select all

LS DRINK C 1/SPIR_TAX log(spir_tax)

fails with the same error while

Code: Select all

LS DRINK C log(spir_tax) 1/SPIR_TAX

produces answers.

[EViews Gareth]

Thanks. We'll look into it.

[EViews Gareth]

Recent patch might fix it.

[bparksb]

I have the same error with a new data set. This time

Code: Select all

ls medal c gdp gdp^2 1/gdp log(gdp)

obtains the error while

Code: Select all

ls medal c gdp gdp^2 log(gdp) 1/gdp

works!
http://econ413.wustl.edu/hinesa-error.wf1

does not. Current support info

=======
PRODUCT
=======
Version: EViews 8 32-bit
Build date: Apr 5 2014
Serial Number: xxxxxxxxx

=========
LICENSING
=========
Standalone xxxxxx: C:\Documents and Settings\All Users\Application Data\IHS EViews\EViews\8.00.000\license.txt

=====
PATHS
=====
Installation path: c:\program files\eviews 8
Working path: G:\E413\ADATA
INI path: c:\documents and settings\xxx\application data\quantitative micro software\eviews
DB INI path: c:\documents and settings\xxx\application data\quantitative micro software\eviews
Help path: c:\program files\eviews 8\help files

===========
ENVIRONMENT
===========
OS: Windows XP Service Pack 3 32Bit
Total Address Space: 1453412k
VMEM Block Space: 1228324k
VMEM Space In Use: 24304k

EViews Address Space: 1206888k
EV Add Space In Use: 2868k
Cores in use: 2

[EViews Glenn]

Good news (sort of).

We've been able to replicate the error. The reason that Gareth wasn't able to see the error is, I believe, due to differences in the global evaluation settings on the relevant machines.

In the dialog for Options/General Options/Auto-series there is a setting for whether to report auto-series evaluation errors in statistical views and procedures. Presumably, bparksb has this set to "Report evaluation errors" and Gareth has it set to "Treat evaluation errors as NAs". EViews is attempting to report the error on bparks' machine as requested, but the error code "1073741824" is not a particularly instructive error message :D. Clearly that has to be fixed, and we'll look at it.

For now, if bparks changes the Global Options setting to Gareth's, the estimation will go through, dropping the missings as required. We'll get the error message cleaned up. Thanks for your patience on this one. It wasn't an easy one to replicate.

[bparksb]

THANKS - that did fix it. Bob