topic Re: Find maximum value from year 0 to year t continuously in SAS Programming

Find maximum value from year 0 to year t continuously

trungcva112 — Thu, 04 Jun 2020 00:50:48 GMT

Re: Find maximum value from year 0 to year t continuously

Shmuel — Sun, 10 Dec 2017 11:59:15 GMT

You can use PROC MEANS with option: max = maxATQ and

CLASS company ID;

VAR ATq;

WHERE datadate between <date created> and <current date>;

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Sun, 10 Dec 2017 12:33:35 GMT

Thank you, I tried but it didn't work

My code is:

proc means data=practice;

max(atq)= max_atq;

class GVKEY;

var atq;

where datadate between <date created> and <current date>;

run;

I dont understand the part <date created> and <current date>. What am I supposed to put into <date created> and <current date>? Because each company has different beginning date. Company 1 data begins at 31/3/1983, company 2 begins at 31/3/1982

Re: Find maximum value from year 0 to year t continuously

Shmuel — Sun, 10 Dec 2017 12:51:03 GMT

Assuming company created date is the lowest value of datadate and current date is the latest one - then you can

just ignore and cancel the where clause.

Did the code worked after removing the where statement ?

if negative - explain what happend and post the log.

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Sun, 10 Dec 2017 13:25:46 GMT

It still doesnt work when I remove the where statement

The log file below:

data practice;
set 'c:\mysaslib\firms.sas7bdat';

NOTE: There were 1522473 observations read from the data set c:\mysaslib\firms.sas7bdat.
NOTE: The data set WORK.PRACTICE has 1522473 observations and 25 variables.
NOTE: DATA statement used (Total process time):
      real time           0.35 seconds
      cpu time            0.35 seconds

138 proc means data=practice;
139 max(atq)=max_atq;
     ---
     180
ERROR 180-322: Statement is not valid or it is used out of proper order.
140 class GVKEY;
141 var atq;
142 run;

NOTE: The SAS System stopped processing this step because of errors.
NOTE: PROCEDURE MEANS used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds

Re: Find maximum value from year 0 to year t continuously

Shmuel — Sun, 10 Dec 2017 13:54:54 GMT

Try next code with the right syntax:

proc means data=practice  ;
  output out=want    max(atq)=max_atq;
  class GVKEY;
  var atq;
run;

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Sun, 10 Dec 2017 14:17:23 GMT

Hi. It works but it produces the maximum value of each company for the entire date range, which is not what I need to do

So the problem is: for each company i:

at t=1: I need to find the maximum value of ATQ from t=0 to t=1

at t=2: Max ATQ from t=0 to t=2

at t=3: Max ATQ from t=0 to t=3

and so on .....

at t=T: Max ATQ from t=0 to t=T

So basically the date range (to find max value) expands when time progress. And I need to find different maximum values of ATQ corresponding with different time range (0 to 1, 0 to 2, 0 to 3, ...... , 0 to T)

I think I need to do some kind of loop?. But I can not find any information on how to do it

Re: Find maximum value from year 0 to year t continuously

Shmuel — Sun, 10 Dec 2017 14:32:37 GMT

What do you mean by t= ?

Is it - first year, first 2 years, first 3 years etc ?

If positive then the use of a data step will be more usefull then proc means:

proc sort data=practice; by GVKEY datadate; run;
data temp;
        year = year(datadate);
run;
data want;
 set temp;
   by gvkey year;
        retain maxATQ;
        if first.gvkey then maxATQ = ATQ; 
        maxATQ = max(maxATQ, ATQ);
       if last.year then output;
run;

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Sun, 10 Dec 2017 14:48:58 GMT

Yes, that's exacly what I mean.

I need to find maximum value in the first year, first 2 years, first 3 years .... and finally the entire time range.

Thank you for giving me some clue. I will try it and give you some feedback as soon as possible

Re: Find maximum value from year 0 to year t continuously

Astounding — Sun, 10 Dec 2017 18:20:07 GMT

It's a little puzzling what you are looking for, since you don't actually illustrate what the outcome should be. You describe the problem in terms of years. But the date values you use to illustrate are quarters.

Since your data set is in order by GVKEY DATADATE, here is a program that may help.

data want;

set have;

by gvkey;

if first.gvkey then max_atq = atq;

else max_atq = max(max_atq, atq);

retain max_atq;

run;

Note that you can have many observations per DATADATE. But the MAX_ATQ value is the max found so far. Therefore, it can be different for two observations having the same DATADATE.

This program will move you in the right direction. To make the requirements clearer, you will need to provide easier to read data, and a sample of what you picture the result to be (especially when you have two identical DATADATE values for the same GVKEY). Then you can get a better suggested solution.

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Mon, 11 Dec 2017 05:00:25 GMT

Hi astounding and shmuel. Your code works.

But Now I have to calculate standard deviation of variable X from quarter I/1983 to quarter II/1983, and then quarter I/1983 to quarter III/1983, and so on.....

It just like before:

at t=1, I need to find standard deviation of X from t=0 to t=1

at t=2, std of X from t=0 to t=2

and so on,

Do you have any idea how to do this? Because the previous code could not applied in this case

Re: Find maximum value from year 0 to year t continuously

trungcva112 — Mon, 11 Dec 2017 13:27:04 GMT

I just posted a sample of my dataset. So could anyone help me please?

Re: Find maximum value from year 0 to year t continuously

Shmuel — Mon, 11 Dec 2017 14:21:33 GMT

Maybe there are some economical or statistical procedures that I'm not familiar with.

You can use the same logic as in computing the maxATQ in order to calculate the standard deviation

using its formula which I dont remeber exactly, that is something like:

- retain sum_x - sum of variable x

sum_sqr_x - sum the squares of variable x

- then compute the STD using those sums retained and _N_ as cumulative number of observations

If you have one observation per qurter then just output each with the last computed values.

if you want comulative per whole years - output it on last.year - as in my post for maxATQ.

Re: Find maximum value from year 0 to year t continuously

Peter_C — Mon, 11 Dec 2017 16:24:39 GMT

you might adapt a solution I offered earlier https://communities.sas.com/t5/General-SAS-Programming/Estimate-standard-deviation-quarter-by-quarter/td-p/420065/jump-to/first-unread-message

That delivered progressively, the STD() for a number stream within each company, over time

Looks like your task is quite similar

Re: Find maximum value from year 0 to year t continuously

Peter_C — Mon, 11 Dec 2017 16:26:03 GMT

https://communities.sas.com/t5/General-SAS-Programming/Estimate-standard-deviation-quarter-by-quarter/td-p/420065/jump-to/first-unread-message

Re: Find maximum value from year 0 to year t continuously

Astounding — Mon, 11 Dec 2017 18:49:03 GMT

The mean can be calculated by keeping track of the sum and N, accumulating as you go. In similar fashion, standard deviation can be written as a function of N, sum, and sum of squared values. If you accumulate all three pieces, you can use the same approach and calculate standard deviation at each point along the way.

Re: Find maximum value from year 0 to year t continuously

uopsouthpaw — Fri, 07 Aug 2020 22:52:17 GMT

Thanks Shmuel,

absolutely helped me with my problem of continuous updating maximums over time.

Wes

Re: Find maximum value from year 0 to year t continuously

mkeintz — Sat, 08 Aug 2020 04:55:20 GMT

For each GVKEY you want maximum ATQ for the entire history of the company through the current DATADATE. Since you are using ATQ, you must be using Compustat quarterly data, correct?

Your data is probably already sorted by gvkey/datadate. If so, then

data want;
  set practice;
  by gvkey datadate;
  retain max_atq;
  if first.gvkey then max_atq=atq;
  else max_atq=max(max_atq,atq);
run;

I use "by gvkey datadate", even though I only need "by gvkey". I threw in the datadate variable just to obligate SAS to report if data are out of order.

For almost all of your results, the DATADATEs within a gvkey will be separated by exactly 3 months, based on whatever month the company uses to end its fiscal year. HOWEVER, you can sometimes get shorter intervals, when a company changes its fiscal year.