Slow Database Refreshing when Workfile Loads

[tchaithonov]

Hi Guys,

A quick question regarding database auto refreshing while workfile that are linked (symbolically, i.e., with a lot of "db::var_abc" in group objects) to the said database is loading. It used to have a pop-up window asking if the user wants to refresh data from the database; lately I haven't seen that at all. I wonder if it's due to my selection of "automatic search database" or "include at start of search order" in the database registry window. It takes me 1 min max to dbopen the database and show all data, but 45 mins to load the workfile (around 2MB), which sounds a bit extreme. Would you please let me know how I could stop EViews from refreshing the database data automatically every time it opens up a workfile? Thanks.

Tchaithonov

[EViews Gareth]

Few things to discuss here.

The pop-up window asking if you want to refresh the data will only occur if you have an explicit data-base link series in the workfile, i.e. a series object that is specified as a data-base link (and has a pink icon). Any implicit links that are in groups, formulas or equations, etc..., will not offer that option and will be refreshed automatically upon opening the workfile.

If you want the behaviour where you can choose to refresh the link or not, create data-base link series in your workfile for each of the series, then include those series links inside your groups etc...

"Automatic search database" and "include at start of search order" should be irrelevant in this case. The only time they would come into play is if you have them turned on, and you have something in the workfile that is mis-specified. e.g. you have an equation object whose dependent variable is "Y", and you don't have a series called "Y" in your workfile.

45 minutes does seem to be an incredibly long time to open a workfile (I can't recall ever seeing one that long). Where is the database stored? Do you have frml objects (or groups involving complicated expressions of the data-base series)? The 1 minute to do the dbopen also seems like a long time too.

[tchaithonov]

Hey Gareth....There are live data transformations within group objects (3,000+ variables), so I guess that's the main reason why it's opening so long. dbopen doesn't take more than 2 secs but Show All takes about a min or so. It's a panel dataset involving 26 cross sections (13,000 data points per variables) and I bet that's the reason why it is opening so slowly. The database is stored locally so it is not a network problem. If I intentionally create one fetch(link) then perhaps I could stop the workfile from refreshing every time?

[EViews Gareth]

No, it would only stop that single series from being updated.

[tchaithonov]

So, I guess, if I don't want the file to open this slowly, I might have to save the data transformations within the groups somewhere else and put them in after the workfile is opened....