About Mscarboncopy

Mscarboncopy · ‎05-08-2019

I agree... but it was not my decision. So now I have 2 data sets - one with the variables a b c d and e and another with X (and key values all in X representing ABCDE) And X is supposed to be numeric, so I can't change it to a string. I think I will do what you suggested, create a number for the combinations AB - AC - AD - AE then BC BD BE and CD CE and DE and all the other combinations as I gave in my example. Thank you !

Mscarboncopy · ‎05-07-2019

Hi I have to rename 5 variables A B C D E (that were selected yes = 1 no = 0) to become one variable X with key values of 1-5 BUT we can have multiple values. So X can be a 1 or 1,2 or 2 or 1,2,3 or 3 or 1,2,3,4 or 4 or 1,2,3,4,5 or 5. So I thought of doing this: if A =1 then X =1; if B= 1 then X =2; if C =1 then X =3; if D then X =4; if E= 1 then X =5; else if (A =1 and B = 1) then X=1,2; if (A =1 and B = 1 C =1) then X=1,2,3; if (A =1 and B= 1 and C =1 and D=1) then X =1,2,3,4; if (A =1 and B = 1 and C=1 and D=1 and E= 1) then X=1,2,3,4,5; Run; but this does not work. Error message: Expecting an arithmetic operator (where the commas were added) I changed the comma with the word and. It "works" but it keeps only one of the values. so in if (A =1 and B= 1 and C =1 and D=1) then X =1 and 2, and 3 and 4; It keeps only one value. Should I have an array instead ? And if so, how can I tell sas to keep all the values when needed ? Thank you in advance.

Mscarboncopy · ‎01-29-2019

Will do. Thank you !!!!

Mscarboncopy · ‎01-29-2019

Thank you for the clarification.

Mscarboncopy · ‎01-29-2019

Good point, however, I want the max value not necessarily the last entry in the case of an error - so I thought that the max value was a good idea, since I need to know what was the last date (as in the most recent, i.e., max). The last entry should have been the most recent date (max) but might not have been (in case of an error). Thanks again,

Mscarboncopy · ‎01-29-2019

data want; set have; array t(*) date:; last_date=t(n(of t(*))); format last_date mmddyy10.; run; I don't know. I don't understand the array, shouldn't the array be all of the date variables ? Someone else suggested this and I can understand it a little better this way: array dates {175} date1-date175; lastdate = max(of dates{*}); Would this be the same idea of what you were thinking? I have not been able to try it yet but will do asap. Thanks

Mscarboncopy · ‎01-29-2019

No gaps and the dates should be in chronological order but yes there could be errors (I was not thinking about that) where one date might be out of chronological order. But those should not be the norm, So a few cases could have an instance where the last date is NOT the actual last date because it could have been coded out of chronological order. Thanks

Mscarboncopy · ‎01-29-2019

It looks something like this but I have 175 dates and 200 + ids. I need to be able to have the last date filled out for each ID - thus I thought of creating a variable X that would be the last date for each, as you can see the last date can be anywhere. ID date1 date2 date3 date4 date5 date6 date7 date8 date9 date10 date11 1 1/1/200 2/2/2006 2 1/2/2003 2/2/2005 3/3/2008 10/10/2009 3 2/6/2007 3/3/2008 10/10/2008 11/11/2008 12/19/2010 12/20/2011 4 3/7/2008 4/4/2009 5/5/2010 5 5/5/2010 6/6/2011 10/10/2008 11/11/2008 12/19/2010 12/20/2011 1/1/2012 3/3/2012 5/5/2013 6 6/6/2011 1/1/2012 3/5/2013 7 1/1/2005 2/2/2005 3/3/2008 10/10/2009 9/9/2010 8 10/3/2000 11/13/2001 6/16/2005 7/7/2008 12/19/2010 12/20/2011 1/1/2012 3/3/2012 5/5/2013 10/3/2014 11/3/2014 9 11/11/2001 12/19/2011 6/7/2012 7/14/2012 10 2/2/2006

Mscarboncopy · ‎01-29-2019

I have 175 possible dates of which some might be missing and I need to single out the last date for each ID. For my other variables - not the date variables but the variables I created an array for - I have about 200 variables. Variable Y was one example of these 200 variables. These 200 variables file is a stacked file, so I will have one ID with 5 entries for variable Y I will have another ID with 100 entries for variable Y and another with 50 entries for variable Y and I wanted to be able to have the total without having to sum the up manually. I have about 230 Ids so it would be a lot to do the sum Id by Id and for 200 different variables ! Thanks.

Mscarboncopy · ‎01-29-2019

Hi I have a variable "date" that can be filled out a max of 175 times for each ID, but I get anywhere from 10 to 50 maybe one or two Ids would have 175 dates. I need to be able to run frequencies to find out - for each id- what was the last date filled out. Is there a data step where I can create a variable "X" that would allow me to then run frequencies for the last date to know which date was the last filled out for each ID? I was thinking of the if first then but use if last instead, would that work ? if last.date then X= date; I am thinking this would single out the last date for each ID in one general command. Am I thinking correctly ? I also need to do a similar thing but in this next case I need to sum all of the other variables, but only those that were coded as 1. I will give the example of variable "Y" Variable Y could have been filled out 175 times but I can get anything from 1 to 175 for each of my Ids. I already did an array to be able to single out the ones that were filled out as 1 but now I need to be able to sum them and they will vary per ID. So one Id will have 10 - and another would have 5 etc..It is a stacked file. I need to have a way to sum only the variables that were filled out (=1) and leave the missing out of the sum. To do this for each ID would take too long, as it varies. I would need to know which ones to keep and which ones to leave out of the sum. Is there a general command that would allow me to create a variable that would be the sum of that variable Y ? but only for the ones that were filled out =1? Leaving the missing out? TIA

Mscarboncopy · ‎12-21-2018

Thank you !! I did this and with a few other steps and it worked 🙂

Mscarboncopy · ‎12-21-2018

Data want; Merge Y Z; by id visit_type; Run; This gives me a data set with 13.000 observations. Multiple observations for each ID. Different variables, different ages. Only a couple of ages match from one file to the other. I need to make sure the ones that do match in age are only in one line instead of 2. Variables in Y are not the same as in Z aside from id visit_type and age. but age var has two different names (depending on the file) so it would not overwrite. one is age_yr the other is visit_age. I am trying to do a comparison to see which ones match but then what ? So this way I can merge them but prevent the overwriting of the other variables... so what is not in Y has to be missing and what is not in Z has to be missing. I don't have a data step for that, it is what I am trying to figure out. I hope that helps. Thanks Mari

Mscarboncopy · ‎12-20-2018

yes. Thank you. I changed the merge by id and age and it worked. it is not overwriting. Now I have another issue: I need to be able to merge the data when the age is the same. I am not sure if I can explain but I will try. When I merge by 2 variables if prevents the overwrite and this was possible because I was talking about it with you here 🙂 But if I have data for the same age for example say I have measurements (from the one file) and diagnoses (from the other) with same age 10. They are in 2 different lines when they should be in one line only. so ID A B C D E F AGE 1 10 5 5 . . . 14 1 . . . 1 1 1 14 How do I tell sas to make this entry only one entry for those cases only? This happens in several instances in the data. Thank you so much

Mscarboncopy · ‎12-20-2018

Thank you. I changed to merge by id and age and it worked. Now I have another issue: I need to be able to merge the data when the age is the same. I am not sure if I can explain but I will try. When I merge by 2 variables if prevents the overwrite. But if I have data for the same age for example I have measurements and diagnoses with age 10. They are in 2 different lines when they should be in one line only. so ID A B C D E F AGE 1 10 5 5 . . . 14 ID 1 . . . 1 1 1 14 How do I tell sas to make this entry only one entry ? This happens in several instances in the data. Thank you so much

Mscarboncopy · ‎12-20-2018

Y has height Body Weight variables Z has diagnoses (URI ASTHMA etc) and drugs variables X needs to have all variables from Y and Z. It is also longitudinal so one ID may have 5 entries in File Y and 10 in File Z. another ID might have 2 entries in File Y and 5 entries in file Z. All different ages. The variable that they both have in common is AGE and I found a way to prevent the ages from overwriting. So if I have one entry from Z and 10 from Y for ID 1- I have 11 entries with different ages for ID 1 but then all of the diagnosis that were 1 in 1 ENTRY are repeated as 1 in all other 10 entries even though it should have been 1 only in ONE ENTRY and 0 in all the other 10 (from Y). I thought they would be missing and then I could tell sas to do if dx =. then dx=0. for example. But with all diagnoses = 1 I don't have a way to give a command to transform it to 0. It will transform others (that must remain 1 to 0) - it will overwrite my correct data in Z. Every ID is different: different ages and different amount of entries and diagnoses.

Online Status	Offline
Date Last Visited	‎08-29-2024 02:05 PM

Re: How to create a nice graph with age group and gender

Re: How to create a nice graph with age group and gender

Re: How to create a nice graph with age group and gender

How to create a nice graph with age group and gender

Re: Creating a new variable based on combination of values in another ...

Re: Creating a new variable based on combination of values in another ...

Creating a new variable based on combination of values in another vari...

Re: flagging same dates across different date var

Re: flagging same dates across different date var

Re: flagging same dates across different date var

Re: How to create a nice graph with age group and gender

Re: How to create a nice graph with age group and gender

Re: Transpose and rename with conditions

Re: Transpose and rename with conditions

Re: Removing repeated values not using first.id

Re: how to rename a numeric variable with multiple choice answers ?

how to rename a numeric variable with multiple choice answers ?

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

Re: How to create a variable based on last variable filled out and sum...

How to create a variable based on last variable filled out and summing...

Re: How to prevent overwriting on a merge

Re: How to prevent overwriting on a merge

Re: How to prevent overwriting on a merge

Re: How to prevent overwriting on a merge

Re: How to prevent overwriting on a merge