The RATS Software Forum

Posted: **Fri Aug 01, 2014 7:32 am**

Hi,

So possibly another strange/terribly ignorant question, but is there a way to drop variables from a set of created data? The code I am using is below.

In short, I have read 14 variables from an excel file into RATS and then created a whole bunch of other variables, most of which I have done so by using loops. There are a few that I do not want to keep, and I tried the DELETE command, but this did not work. I realise this is because it works on saved data files only and my created data is effectively in a "RAM" version of the dataset. Is there a solution that I have missed?

Many thanks.

Regards
Siobhan Redford

Code: Select all

OPEN DATA "C:\Users\p174343\PhD\Yield curve\CFS\Monthlydata.xlsx"
CALENDAR(m) 1986:01:31
DATA(FORMAT=XLSX,ORG=COLUMNS) 1986:01:31 2014:06:30 TB3 JAYC1 JAYC2 JAYC3 JAYC4 JAYC5 JAYC6 JAYC7 $
JAYC8 JAYC9 JAYC10 JAYC15 JAYC20 JAYC25

dofor i = TB3 to JAYC25
set %s("D3"+%l(i)) = (i{3})-(i{0})
end dofor i

dofor i = JAYC1 to JAYC25
set %s("D12"+%l(i)) = (i{12})-(i{0})
end dofor i

dofor i = JAYC4 to JAYC20
set %s("D24"+%l(i)) = (i{24})-(i{0})
end dofor i

dofor i = JAYC10 to JAYC25
set %s("D60"+%l(i)) = (i{60})-(i{0})
end dofor i

dofor i = JAYC1 to JAYC25
set %s("S3"+%l(i)) = (i{0})-(TB3{0})
end dofor i

dofor i = JAYC2 to JAYC25
set %s("S12"+%l(i)) = (i{0})-(JAYC1{0})
end dofor i

dofor i = JAYC4 to JAYC20
set %s("S24"+%l(i)) = (i{0})-(JAYC2{0})
end dofor i

delete D24JAYC5 S24JAYC5 D24JAYC7 S24JAYC7 D24JAYC9 S24JAYC9 D24JAYC15 S24JAYC15

dofor i = JAYC10 to JAYC25
set %s("S60"+%l(i)) = (i{0})-(JAYC5{0})
end dofor i

Posted: **Fri Aug 01, 2014 8:09 am**

Can't you just ignore them? That's not a data set that's going to produce any memory shortages.

Posted: **Fri Aug 01, 2014 9:06 am**

I could, but it would create an additional problem in which I then have to scale the variables using a particular formula, which I haven't managed to figure out how to code, especially since the end variables are not consecutive and then use those for the regressions. By not having the variables I want to drop, I can then run the regressions I will be running using these created variables using a loop as everything is then in order.

This is the set of variables I create next :

Code: Select all

set AS3JAYC1 = S3JAYC1/3
set AS3JAYC2 = S3JAYC2/7
set AS3JAYC3 = S3JAYC3/11
set AS3JAYC4 = S3JAYC4/15
set AS3JAYC5 = S3JAYC5/19
set AS3JAYC6 = S3JAYC6/23
set AS3JAYC7 = S3JAYC7/27
set AS3JAYC8 = S3JAYC8/31
set AS3JAYC9 = S3JAYC9/35
set AS3JAYC10 = S3JAYC10/39
set AS3JAYC15 = S3JAYC15/59
set AS3JAYC20 = S3JAYC20/79
set AS3JAYC25 = S3JAYC25/99
set AS12JAYC2 = S12JAYC2
set AS12JAYC3 = S12JAYC3/2
set AS12JAYC4 = S12JAYC4/3
set AS12JAYC5 = S12JAYC5/4
set AS12JAYC6 = S12JAYC6/5
set AS12JAYC7 = S12JAYC7/6
set AS12JAYC8 = S12JAYC8/7
set AS12JAYC9 = S12JAYC9/8
set AS12JAYC10 = S12JAYC10/9
set AS12JAYC15 = S12JAYC15/14
set AS12JAYC20 = S12JAYC20/19
set AS12JAYC25 = S12JAYC25/24
set AS24JAYC4 = S24JAYC4
set AS24JAYC6 = S24JAYC6/2
set AS24JAYC8 = S24JAYC8/3
set AS24JAYC10 = S24JAYC10/4
set AS24JAYC20 = S24JAYC20/9
set AS60JAYC10 = S60JAYC10
set AS60JAYC15 = S60JAYC15/2
set AS60JAYC20 = S60JAYC20/3
set AS60JAYC25 = S60JAYC25/4

Posted: **Fri Aug 01, 2014 10:55 am**

Sorry. I'm trying to figure out the pattern for those, but I'm not sure I see it. What are the conventions in the names?

Posted: **Fri Aug 01, 2014 1:26 pm**

They are yields on zero coupon bonds of maturity 1, 2, 3 years etc as indicated by the numbers at the end of the name. I am adjusting spreads similar to Campbell and Shiller 1991 (Yield spreads and interest movements: a bird's eye view, Review of Economic Studies, v 58, p 495 - 514). Basically they are being divided by (1/time to maturity - 1), in months, it would be easier to write a loop if I didn't skip years from 10 to 25. Also, the requirement is that the maturity of the short yield is divides exactly into the long yield. Which is why I would like to drop the non-integer yields. Does this make some sense?

Posted: **Fri Aug 01, 2014 2:04 pm**

Code: Select all

set AS3JAYC1 = S3JAYC1/3
set AS3JAYC2 = S3JAYC2/7
set AS3JAYC3 = S3JAYC3/11
set AS3JAYC4 = S3JAYC4/15
set AS3JAYC5 = S3JAYC5/19
set AS3JAYC6 = S3JAYC6/23
set AS3JAYC7 = S3JAYC7/27
set AS3JAYC8 = S3JAYC8/31
set AS3JAYC9 = S3JAYC9/35
set AS3JAYC10 = S3JAYC10/39
set AS3JAYC15 = S3JAYC15/59
set AS3JAYC20 = S3JAYC20/79
set AS3JAYC25 = S3JAYC25/99

can be handled with a DOFOR loop across the term:

dofor term = 1 to 10 15 20 25
set %s("AS3JAYC"+term) = %s("S3JAYC"+term)/(4*term-1)
end dofor term

It looks like something similar with work with the others.

The RATS Software Forum

Dropping created variables from the "current dataset"

Dropping created variables from the "current dataset"

Re: Dropping created variables from the "current dataset"

Re: Dropping created variables from the "current dataset"

Re: Dropping created variables from the "current dataset"

Re: Dropping created variables from the "current dataset"

Re: Dropping created variables from the "current dataset"