About ballardw

ballardw · ‎12-12-2011

It looks like you are missing X, Y, and/or Z values on a couple of lines where you are calling the LABEL function. Look at the 7th obseravation (record) in your annotate dataset.

ballardw · ‎12-12-2011

On potential source of this behavior is if the datasets with the counts shown come from PROC MEANS or SUMMARY output with the age as a CLASS variable with a grouping format as implied by your example, the actual value kept in the datasets could be a different numeric value. Suppose the lowest age in one set before PROC MEANS is 13 and the in the other set the lowest is 15 the actual values in the sets would be 13 and 15 and when merged 13 does not equal 15 so would be on a different line. The other categories may not show this behavior if they all had the same minimum values per age group. If not, you need to let us know how you are building sets 1 and 2.

ballardw · ‎12-08-2011

If available to you, PROC GPLOT using a creative AXIS statement listing values of the days with data might work. You would get a warning in the log about non-uniform values on your ticks but might accomplish what you want.

ballardw · ‎12-08-2011

Thank you for this information. I was really hoping to find something that ODS addresses that would let me change the text displayed as well as how to access the object that contains the number of missing. The advantage of creating a bunch of one-way tables with a single pass of the data isn't going to be there apparently. The MISSPRINT table option doesn't quite work as I wanted to display the missing values somewhat separated. I was hoping to find something like the FMISSING available in the Base.Freq.CrossTabFreqs. Maybe I can find a way to make a crosstab look like a one-way table.

ballardw · ‎12-08-2011

If the Config.sas, or as with version 9, SASV9.CFG file actually needs to be modified, then there is something that only takes effect on startup. So you can't actually execute those files in a runnings session. It's been awhile since I played around with some of this but it may be that you want to create an additional shortcut to SAS, or a batch file that starts a SAS session, and point to those files. For examle my Windows desktop icon in the properties shows the following in the Target section. "C:\Program Files\SAS\SASFoundation\9.2(32-bit)\sas.exe" -CONFIG "C:\Program Files\SAS\SASFoundation\9.2(32-bit)\nls\en\SASV9.CFG" If you have a shortcut, copy it and replace the existing similar with the new config file path. Instead of the autoexec parameter you could add the line: -autoexec="S:\GenPeds\GenPeds\JHale\DualData\test_exe.sas" to that configuration file. Note that this form as an equal sign.

ballardw · ‎12-08-2011

I wouldn't be surprised to find that since the SQL server was updated some permissions behave differently or some policy for your login behaves differently with the new software. I ran into oddnes with Access connecting to and MS SQL database that made the fields behave differently. It's probably time to talk with the DB manager about things. should Add that nothing happend with user login on SQL server so the rights for the user are same as before.Only thing I found out is,that SQL server was updated.But as I said nothing happend to login and everything is same as before.

ballardw · ‎12-07-2011

Have you tried specifying the version 6 engine for your version 6 data set? Also there are references to DDE and OS/2. The DDE calls might be significantly different with current versions of Windows.

ballardw · ‎12-06-2011

Goal: Control the text and appearance of the Frequency Missing= message and format of the numeric value displayed at the end of a PROC FREQ one-way table when there are missing values. For example display "Not Tested = 3,422" in red italics. I am using SAS 9.2.3 and have looked through a lot of ODS references. I thought this would require modification to the BASE.FREQ.ONEWAYLIST template but the definition doesn't have anything I can see that would control that part.

ballardw · ‎12-06-2011

I think that can happen when you open a second instance of SAS that both use the same configuration. The first one in effect locks the default SASUSER.PROFILE and the second can't read it hence the messages. The main effect is that any customizations that are stored in SASPROFILE aren't available. If this isn't from a second instance it could mean that the catalog is corrupted. You could try either renaming or deleting the catalog using a system command out side of SAS. When SAS next starts is should rebuild a default profile from SASHELP. That profile will be factory defaults though.

ballardw · ‎12-01-2011

Check Tools>Options>Preferences>Edit to see if your Autosave setting is turned on. If so then it was saved somewhere but it is OS and version dependent. Unfortunately the system option AUTOSAVELOC doesn't report where. I seem to remember finding some code files, with and extension other than .SAS in a Windows TEMP or TMP folder. If I remember correctly they may have had a name like $Autoxxxxx That was in with Windows XP and SAS 9.2.0 neither of which I currently have so I can't check for them.

ballardw · ‎12-01-2011

Let me first put forward my concern. I have multiple users each of who upload an excel file to my sas directory. Post which a batch job runs to process these files Now there are two columns one is an ID and the other a Pincode. Now ideally these two columns are supposed to be character columns with numeric digits. But certain users may upload it as character in their excel where as others may upload it as numeric. So i wanted to check if it was possible to import the data as character irrespective of the type in which the user has uploaded. Since it appears that you have users making a choice then the instructions to them could be modified from uploading the .xls to SAVE AS .CSV. Then what they have as the format in Excel won't matter. This is an old problem. I ran into similar issues with reading Lotus 123 .PRN files into SAS 5 around 1991 with the added joy of getting them onto a mainframe from floppies. That experience is why I avoid spreadsheets as much as possible.

ballardw · ‎11-30-2011

By length here I suspect you are referring to text length of responses. I would try specifying the lengths of any common variables before the Merge or Set statement. IF the variables are the same name you probably want to SET instead of MERGE or you'll only get one. If the variables aren't the same name there is no problem. Data new; length TextVar1 TextVar2 $ 200; set Data1 Data2; etc.

ballardw · ‎11-16-2011

I believe using IMPORT you may have to make the changes in Excel as SAS is requesting data that is processed through the JET or similar Excel engine. The engine tells SAS whether the values are numeric or string. If you do not have a lot of workbooks/sheets to process you could save the worksheets as CSV and use the Proc Import for CSV. This alone doesn't fix your issue. However the proc import will create data step code to read the CSV file and you can change the INFORMAT statements to $ formats of the appropriate length.

ballardw · ‎11-16-2011

Did you notice that changed from a PUT to a PUTC? Also are you wanting to include the record if any of the variables meet the condition or all of the variables meet the condition? If so the loop approach you are using probably isn't going to work. You may need to be a bit more explicit instead of using the array: where [put(diag_cd1,$diagl.) eq 'AIDS'] OR [put(diag_cd2,$diagl.) eq 'AIDS'] or <continue through all variables>; if ALL of the codes then replace the OR with AND.

ballardw · ‎11-16-2011

When I tried a DATETIME16. format with that value of 1636523977 for cdate I get 10NOV11:05:59:37. Does that look like the correct date for your example? If so you want the DATEPART of CDATE.

Online Status	Offline
Date Last Visited	4 hours ago