Hi all,
We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
There are no missing values for long strings whether user-defined or system-defined. Try Autorecode to make your long strings into numbers and then declare the number assigned to the blank string as missing.
________________________________ From: SPSSX(r) Discussion on behalf of Parise, Carol A. Sent: Thu 9/21/2006 6:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
In reply to this post by parisec
Stephen Brand
www.statisticsdoc.com Carol, You can declare a missing value for a string variable as follows. Suppose fname and lname are string variables: missing values fname lname (" ") . will result in fname and lname being treated as missing in frequencies. A good site to read up on this is: http://www.ats.ucla.edu/stat/spss/faq/countstr.htm HTH, Stephen Brand For personalized and professional consultation in statistics and research design, visit www.statisticsdoc.com -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of Parise, Carol A. Sent: Thursday, September 21, 2006 7:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
In reply to this post by parisec
Stephen Brand
www.statisticsdoc.com Carol, As soon as I sent my previous message, I realized that you were asking about long strings, which are a different animal, as you say. My example applies to names of 8 characters or less. My apologies. Since you cannot declare a missing value for long strings, another approach would be to run your frequencies procedure after a temporary select if. temp. select if (lname gt ' ' ). freq var lname. HTH, Stephen Brand For personalized and professional consultation in statistics and research design, visit www.statisticsdoc.com -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of Parise, Carol A. Sent: Thursday, September 21, 2006 7:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
In reply to this post by Beadle, ViAnn
In SPSS 14 or later, you can automatically recode blanks into user-missing values with Autorecode (and Autorecode retains the original string values as value labels for non-blank values):
autorecode longstring /into numvar /blank=missing. ________________________________ From: SPSSX(r) Discussion on behalf of Beadle, ViAnn Sent: Thu 9/21/2006 7:34 PM To: [hidden email] Subject: Re: stuggling with strings There are no missing values for long strings whether user-defined or system-defined. Try Autorecode to make your long strings into numbers and then declare the number assigned to the blank string as missing. ________________________________ From: SPSSX(r) Discussion on behalf of Parise, Carol A. Sent: Thu 9/21/2006 6:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
In reply to this post by parisec
Hello Carol,
Just a note that sometimes using select if function under temp does not work and so with count function. Using autorecode is also a safe method along with reducing the length of string variable and then declare the missing. For e.g Var format (a8) * This will reduce the variable to short string. Missing values varname (missing value) Regards, Manmit -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Statisticsdoc Sent: Friday, September 22, 2006 6:23 AM To: [hidden email] Subject: Re: stuggling with strings Stephen Brand www.statisticsdoc.com Carol, As soon as I sent my previous message, I realized that you were asking about long strings, which are a different animal, as you say. My example applies to names of 8 characters or less. My apologies. Since you cannot declare a missing value for long strings, another approach would be to run your frequencies procedure after a temporary select if. temp. select if (lname gt ' ' ). freq var lname. HTH, Stephen Brand For personalized and professional consultation in statistics and research design, visit www.statisticsdoc.com -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of Parise, Carol A. Sent: Thursday, September 21, 2006 7:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
In reply to this post by parisec
Hi
Have you tried recode into same variable There you can specify system or user missing as system missing Hope that helps Shorful On 9/22/06, Parise, Carol A. <[hidden email]> wrote: > > Hi all, > > We received a file that originated in Excel that contains a bunch of long > string variables which means declaring missing values is not gonna happen. I > didn't think this would be a problem but SPSS is reading blanks as values > rather than reading these blanks as system missing which screws up the > frequency computations. > > Anyone know why this is going on and what the workaround is? > > > > Thanks for your insights. > Carol > |
In reply to this post by parisec
woah! i never used autorecode before. I just tried it on this data set and it worked perfectly. that is one great command that i know i'll be using again.
thanks to all who replied. Carol -----Original Message----- From: Beadle, ViAnn [mailto:[hidden email]] Sent: Thursday, September 21, 2006 5:35 PM To: Parise, Carol A.; [hidden email] Subject: RE: stuggling with strings There are no missing values for long strings whether user-defined or system-defined. Try Autorecode to make your long strings into numbers and then declare the number assigned to the blank string as missing. _____ From: SPSSX(r) Discussion on behalf of Parise, Carol A. Sent: Thu 9/21/2006 6:56 PM To: [hidden email] Subject: stuggling with strings Hi all, We received a file that originated in Excel that contains a bunch of long string variables which means declaring missing values is not gonna happen. I didn't think this would be a problem but SPSS is reading blanks as values rather than reading these blanks as system missing which screws up the frequency computations. Anyone know why this is going on and what the workaround is? Thanks for your insights. Carol |
Free forum by Nabble | Edit this page |