Date: Fri, 20 Jun 1997 13:27:13 -0400
Reply-To: rcoleman@worldnet.att.net
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Ron Coleman <rcoleman@WORLDNET.ATT.NET>
Organization: Links Analytical, Inc.
Subject: Re: Medians in Proc Tabulate
Content-Type: text/plain; charset=us-ascii
Unfortunately, PROC TABULATE does not do medians. You can use PROC
UNIVARIATE for any percentiles including 50th (median).
proc univariate data=SS.LOADS ;
by staid;
var basload totload;
run;
Ron
L. Joseph Bachman wrote:
>
> I have a large data set of the following format:
>
> STA-ID YEAR BASELOAD TOTAL_LOAD, etc.....
>
> 1 1991 34 56
> 1 1992 45 89
> 1 1993 32 67
> 1 1994 42 56
> 2 1972 103 120
> 2 1973 110 130
> 2 1974 140 180
> .
> .
> etc...
>
> I want to make a table that will give me the median values over all years
> for each STA-ID. The closest thing I could figure out to do was use PROC
> TABULATE (Through the tables function in ASSIST):
>
> ---begin included SAS code file------
>
> /*------------------------------------------------------------------*
> | Summary: |
> | This program creates a simple tabular report using the |
> | data set SS.LOADS and displaying across levels of |
> | analysis variables and statistics and a down level of |
> | classification variables. |
> | Generated: 19JUN97 11:14:45 |
> *------------------------------------------------------------------*
> | The OPTIONS statement specifies the dimensions of the printed |
> | output in characters per line and lines per page and whether or |
> | not the current date and page number are printed. |
> | The TITLE and FOOTNOTE statements clear any previously defined |
> | titles or footnotes. |
> *------------------------------------------------------------------*/
> options linesize=132 pagesize=26 date number pageno=1;
> title;
> footnote;
> proc tabulate data=SS.LOADS ;
> table STAID
> ,(BASELOAD TOTLOAD) * ( 'N' 'MEAN')
>
> ;
> var BASELOAD TOTLOAD;
> class STAID;
> run;
>
> -------end included SAS code file ----
>
> This produced output of the form...
>
> ------------------------------------------------------------------------
> | |baseflow nitrate load, in| total nitrate load, in |
> | | tons | tons |
> | |-------------------------+-------------------------+
> | | N | MEAN | N | MEAN |
> |------------------+------------+------------+------------+------------+
> |station ID | | | | |
> |------------------| | | | |
> |01491000 | 8.00| 83.88| 19.00| 120.15|
> |------------------+------------+------------+------------+------------+
> |01493500 | 0.00| .| 18.00| 17.86|
> |------------------+------------+------------+------------+------------+
> |01503000 | 13.00| 1646.18| 16.00| 2695.56|
> |------------------+------------+------------+------------+------------+
> |01509150 | 5.00| 6.98| 6.00| 12.82|
> |------------------+------------+------------+------------+------------+
> |01518000 | 11.00| 74.21| 11.00| 269.55|
> |------------------+------------+------------+------------+------------+
> |01518700 | 14.00| 128.41| 16.00| 319.16|
> |------------------+------------+------------+------------+------------+
> |01520000 | 8.00| 154.81| 21.00| 278.57|
> ------------------------------------------------------------------------
>
> Which is what I want, except for one thing. Proc tabluate appears to have a
> limited list of summary statistics available to it. I had wanted to tabulate
> _medians_ for these stations, but instead the on thing available was _means_.
>
> Is there some other way I can generate such a table, and be able to have
medians
> computed? Also, would it be possible to display the minimum and maximum
years
> for each sation in the same table?
>
> Thanks in advance for yur suggestions.
>
> Joe Bachman
> jbachman@usgs.gov
--
Ron Coleman A SAS Quality Partner.
Links Analytical, Inc. Linking your data to your business!
3545-1 St. Johns Bluff Rd. Suite 300
Jacksonville FL 32224
rcoleman@worldnet.att.net
| 
