Help with importing cross section constant data

[notphilip]

Hi all,

So I have one massive dataset consisting of option valuations. Each entry has greeks, stats, and prices for a strike/expiry/day combination sold on a given day for all combinations available on that given day. This is thus a massive data set with about 800 entries per day over a year, or about 203000 data points total. As a workfile, it is organized by using expiry and strike price as cross sections, and the date for the time series.

I also have multiple smaller datasets consisting of high, low, open, and close prices for the stocks such that each data set focuses on a single stock and has one entry per day over the same time period as the options data, with a total of 254 entries per set. I have ensured that the date formatting matches exactly with the date series for the options data set.

Is there a way to import these smaller sets such that eviews will associate their data with each data point with a matching date, regardless of strikes and expiries? Thus it would expand these smaller series across all 203000 data points with the same exact data for every point with a matching date?

Looking at previously compiled data, it appears that my predecessor either figured out such a way or he must have edited his smaller data sets to contain all the recurring points for all strike/date/expiry combinations. The latter would be extremely time consuming

[EViews Gareth]

If you just copy from the smaller dataset to the bigger one, EViews should automatically expand it (assuming it recognises the dates as being the same).

[notphilip]

If you just copy from the smaller dataset to the bigger one, EViews should automatically expand it (assuming it recognises the dates as being the same).

What may be wrong if it is not recognizing the dates as the same? They are formatted exactly alike and the smaller sets are both in csv and xls form, and I don't think there's anything indicating the date to be a string. But when I import the date series for the small sets as the same name as the large set date series, all entries become n/a. Then eviews seems to generate false strike-expiry combinations for the small set data and associates it with weird or "?" dates.

What may I be doing wrong?

[EViews Gareth]

Hard to tell without seeing the files.

[notphilip]

Hard to tell without seeing the files.

Yeah unfortunately, they are huge and some of the stats are under NDA. I'm glad to know that it is at least user error, and that the data expansion is a feature of eviews. Thanks.

[notphilip]

If you just copy from the smaller dataset to the bigger one, EViews should automatically expand it (assuming it recognises the dates as being the same).

Just to clarify, does this work in Eviews 6? And by copy, do you mean by opening the xls file via import? I still cannot determine why the program is interpreting the dates to be different.

[EViews Gareth]

You might have to explicitly tell it how the dates are structured when you do the copy.

[notphilip]

You might have to explicitly tell it how the dates are structured when you do the copy.

Where/how do I specify it? I have tried the "reset current sample to (I)current sample (2) workfile range (3) To end of range" options, and it seemed to do the same thing every time. I feel there may be something I should modify in import sample box. By default, it has the correct range of dates "9/09/2010 9/09/2011."

If it helps, the workfile that I am importing to has the following structure:
Range: 09/09/2010 09/09/2011 x2730 Dim(227, 22, 251)
Sample: 09/09/2010 09/09/2011 -- 202812 obs


Thanks a ton! My research adviser has not been able to assist me with this at all, so your help has been crucial. With a million deadlines coming up, any advice to help save me from wasting time fiddling around is incredibly valuable.

[EViews Gareth]

You should open up the smaller dataset as a new workfile in EViews. Then copy from that smaller one to the larger one. When you do the copy you can use Paste-Special (rather than Paste) and set the copy stuff.

[notphilip]

You should open up the smaller dataset as a new workfile in EViews. Then copy from that smaller one to the larger one. When you do the copy you can use Paste-Special (rather than Paste) and set the copy stuff.

Holy crap that worked perfectly! You are my eff'ing hero! You have no idea how many ways I tried to attempt this otherwise haha.