Export group to Excel

[Phil09]

I constructed a group object from panel data objects. When I export the group to Excel with the following statement, the observation ID (date+cross section) is missing. How do I change that?

wfsave(type=excelxml) {%tableloc}state_backtest.xlsx range="forecast!b1" @smpl {%graph_sample} @keep __g_backtest

[EViews Gareth]

Put the id series into the group.

When you tell EViews to just keep a group it literally only keeps the series in that group.

[Phil09]

If the reason behind export failing to include observation ID is because the group does not retain it, then the following statement should do it:

wfsave(type=excelxml) {%tableloc}state_backtest.xlsx range="actual!a1" @keep {%varlist} @smpl {%graph_sample}

where %varlist is a list of panel data series. But this statement did not produce observation ID either in the Excel table either. The problem seems to lie with exporting panel data to Excel.

[EViews Gareth]

Code: Select all

wfcreate a 1990 2000 10 alpha mycrosssections = "myid" + @str(@crossid) alpha mydates = @datestr(@date, "YYYY") series y=@nrnd group g mydates mycrosssections y wfsave(type=excelxml) myfile.xlsx @keep g

Works just fine.

[Phil09]

Great. It seems to me that I need to add the date and the cross section explicitly to the group.

Given a panel data series, how do I extract its date and cross section?

[EViews Gareth]

Code: Select all

alpha mycrosssections = "myid" + @str(@crossid) alpha mydates = @datestr(@date, "YYYY")

or @pageids

[Phil09]

My panel data series is indexed by state abbreviation and date such as "AK - 50Q1". @crossid gives me index 1, ..., 50, rather than state abbreviation. How do I obtain the latter?

@PAGEIDS gives an error message: @PAGEIDS is not a Genr or series expression function.

[EViews Gareth]

I'm a bit confused - do you know the name the of your cross-section identifier?

[Phil09]

The name of my cross section id is _P. For instance if I want to select a sample from the state of Alaska, I issue the following statement:
smpl 2008 2010 if _P="AK"

Eviews documentation on series object does not contain a lot of information on panel data series. Can you point me to some good reference?

[Phil09]

I also wish to know how to extract a particular data point from my panel data series X. For instance, I wish to assign the value of the series value at 2010Q2 for the cross section of "AK" to a scalar. How do I reference that data point?

[Phil09]

For a series without cross section, I can obtain the value at any date with @elem(X, "2010Q2"). When the series is a panel, I get the error message: "Date does not uniquely define observation in panels".

[EViews Gareth]

The name of my cross section id is _P. For instance if I want to select a sample from the state of Alaska, I issue the following statement:
smpl 2008 2010 if _P="AK"

So just include _p in the group?

[Phil09]

Gareth, thank you. That worked!

My mistake is that I kept on thinking that cross section ID is attribute of a panel data series. It is actually an object of the panel data workfile, in which all series have the same cross section ID.

[EViews Gareth]

I also wish to know how to extract a particular data point from my panel data series X. For instance, I wish to assign the value of the series value at 2010Q2 for the cross section of "AK" to a scalar. How do I reference that data point?

There is no @elem for panels, unfortunately. You'll have to set the sample to the observation you want, then use @max or something to extract the value.