Eviews 13 bugs - @rowextract

[leocaza]

Hi,

First of all, thank you for being there to help us, your help in general is often taken for granted, but it's precious and should be perceived otherwise.

There is a lot of instability in the Eviews 13 version at the moment, several processes no longer work due to bugs (corrupted excel, matrix function no longer working)... Here in my company, we will use only Eviews 12 again for the time being, while Eviews 13 is more stable, we can't work with such an unstatble environment (this is the case for many licenses). It is the first time we have trouble to run all our process after an Update.

Here are the bugs:

**********************************************
**********************************************
Bug #1
We are running into some issue with Eviews 13 and are wondering if you knew about some of the bug. One we have clearly identify is a bug with the function @rowextract when working with matrix. One of our code that was running fine with Eviews 12 is not working anymore with Eviews 13 (Oct 31 2022 build). I wrote a small script to replicate the issue 

(page=Q)Q 2010 @now 

matrix(10,10) m1
nrnd(m1) 

VECTOR ER = @rowextract(m1,2) 

In Eviews 12, the function will return a vector with the correct value for the row that is extract. In Eviews 13, I get this error and the function return a vector of length 1 with a 0. 
error : A vector cannot be resized to have more tham 1 column in "Vector ER = @rowextract(M1,2)"
**********************************************
**********************************************
Bug #2
The command to save data directly in excel, which worked well in eviews 12, now does not work correctly in eviews 13, because it corrupts the excel file (when opening the file).
Ex: wfsave(type=excelxml, mode=update) "path" range="Page" byrow @keep variables @smpl {%Date} {%Date}
**********************************************
**********************************************



Remember this : we can't update Eviews as easily as you think, so the first version needs to be "impeccable" (I know that it's probably not the developers's fault, but some really bad management planning (as usual)). I pity you
Regards,
Leo

[EViews Gareth]

Hi. Thanks for the nice messages.

Bug #1 isn't really a bug, but a stricter interpretation of objects in EViews, to allow better programming control. Specifically, a VECTOR object in EViews has a single column and multiple rows. The rowextract function will extract a single row of a matrix, which would result in multiple columns and a single row. Hence the error message.

You would either need to change to using a ROWVECTOR, or take the transpose to store in a VECTOR.


Do you have an example workfile that results in a corrupt .xlsx with the command used in Bug #2?

[leocaza]

I'll try to get the workfile, but it doesn't come from me, it won't be easy on a short time. I'll come back later. Thank you!