Prev: Lookup table from excel spreadsheet
Next: PROC MIXED/GLIMMIX disagreement
From: astrogirl on 17 Jun 2010 15:22
I’m running SAS 9.2 with Windows Vista and I have run across an issue
while importing SPSS files (I'm using version 18) into SAS. I have
tried to import them as .dbf, .sav, and .por and I get the same issue
every time. The files quadruple or quintuple in size from the SPSS to
the SAS format. I have one file in particular that I use most often
(it’s updated each month) and currently in SPSS format it’s 30GB. When
I import it into SAS each of the ways listed above, it becomes around
150 GB. It still has the same number of variables and observations as
in the various SPSS formats but the size has grown by five times.

Usually I work in SPSS but I’m trying to learn SAS because I can
process large files more rapidly that way. So normally this file I get
is in text format and another person in my department imports them
into SPSS and sets them up as .sav files because that is the program
that most of us use and there are several of us that use that
particular file regularly. It won’t import into SAS however as a text
file because some of the fields are character variables and some have
issues. One in particular, last name, is troublesome because it has
spaces between the letters and when it’s importing in SAS, it hits
those spaces and stops. I know I can remove the spaces but doing
anything with these files in SPSS takes hours because they are so
large so I would like to avoid any extra steps like that if possible
as I am trying to be more efficient.

Ideally, I would like to be able to import the .sav file into SAS and
have it be around the same size as the original. Is that possible? If
not I would appreciate any other import suggestions to keep the file
size down.

Thanks,
Angie
From: data _null_; on 17 Jun 2010 16:02
On Jun 17, 2:22 pm, astrogirl <servo262...(a)yahoo.com> wrote:
> It won’t import into SAS however as a text
> file because some of the fields are character variables and some have
> issues. One in particular, last name, is troublesome because it has
> spaces between the letters and when it’s importing in SAS, it hits
> those spaces and stops.

You comment about not being able to "import" the file into SAS seem
odd. I assume by IMPORT you mean read the field in records from a
text file. If you can read the file into SPSS you can surely do the
same with SAS. I expect you have make a simple mistake that can be
easily solved.

So forget about SPSS and read the file directly into SAS. Post an
example of the records from the file and ask has how best to read the
data.

From: astrogirl on 1 Jul 2010 10:56
Thanks for the responses and I’m sorry it’s taken me so long to
respond but I got sidelined with another project.


Here’s the import program that I set up going through the wizard using
a tab delimited text file format. My file doesn’t have variable
names.


PROC IMPORT OUT= CM_FILES.test
DATAFILE= "N:\CM files\cm_enrollment_instance.txt"
DBMS=TAB REPLACE;
GETNAMES=No;
DATAROW=1;
RUN;


Here’s a sample of my data and this includes the record where I’m
having the problem. I can read the tab delimited text file into SAS up
until it hits the record with the last name ‘O NEAL’ then it stops.


44 0 1234567 SPEARS CARRIE
50 0 1234567 SPEARS CARRIE
1 0 .
43 0 .
1 0 1234567 SANCHEZ ISMAIL
1 0 1234567 SANCHEZ ISMAIL
98 0 1234567 ARNOLD JR JOHN
1 0 1234567 ARNOLD JR JOHN
21 0 1234567 ARNOLD JR JOHN
58 1 1234567 SMITH-KONES AMY
25 1 1234567 SMITH-KONES AMY
1 1 1234567 SMITH-KONES AMY
1 1 1234567 Garrett Thomas
V
69 1 1234567 O NEAL COLLEEN
1 1 1234567 O NEAL COLLEEN
5 1 1234567 FULLER ALI
S
2 1 1234567 CALHOUN SLY
M
70 1 1234567 CALHOUN SLY
M
16 1 1234567 DANIEL JR. MICHAEL
J
1 1 1234567 DANIEL JR. MICHAEL
J
1 1 1234567 DANIEL JR. MICHAEL
J
204 1 1234567 KURCZEWSKI-LYONS DAWN
M

Here’s some info about the fields shown in the sample data set.

cm_days_in_service = numeric 10 spaces
cm_valid_ssn_flag = numeric 1 space
customer_id = numeric 10 spaces
last_name = character 30 spaces
first_name = character 20 spaces
middle_initial = character 1 space

Thanks!

Angie
From: Tom Abernathy on 3 Jul 2010 00:55
Hard to tell with the mangling of the example text by the email
system. Doesn't look tab delimited in the email.
You do not need to use PROC IMPORT to read a text file. Just write
your own DATA step.
Especially if you already know the fields.

data cm_files.test ;
infile'.....' truncover dlm='09'x dsd ;
length
cm_days_in_service 8
cm_valid_ssn_flag $1
customer_id $10
last_name $30
first_name $20
middle_initial $1
;
input cm_days_in_service -- middle_initial ;
run;

Note that SAS stores numbers as floating point. So your 10 digit
number will be stored in a normal 8 byte long numeric variable.


On Jul 1, 10:56 am, astrogirl <servo262...(a)yahoo.com> wrote:
> Thanks for the responses and I’m sorry it’s taken me so long to
> respond but I got sidelined with another project.
>
> Here’s the import program that I set up going through the wizard using
> a tab delimited text file format. My file doesn’t have variable
> names.
>
> PROC IMPORT OUT= CM_FILES.test
>             DATAFILE= "N:\CM files\cm_enrollment_instance.txt"
>             DBMS=TAB REPLACE;
>      GETNAMES=No;
>      DATAROW=1;
> RUN;
>
> Here’s a sample of my data and this includes the record where I’m
> having the problem. I can read the tab delimited text file into SAS up
> until it hits the record with the last name ‘O NEAL’ then it stops.
>
> 44      0       1234567 SPEARS                                  CARRIE
> 50      0       1234567 SPEARS                                  CARRIE
> 1       0       .
> 43      0       .
> 1       0       1234567 SANCHEZ                                 ISMAIL
> 1       0       1234567 SANCHEZ                                 ISMAIL
> 98      0       1234567 ARNOLD JR                               JOHN
> 1       0       1234567 ARNOLD JR                               JOHN
> 21      0       1234567 ARNOLD JR                               JOHN
> 58      1       1234567 SMITH-KONES                             AMY
> 25      1       1234567 SMITH-KONES                             AMY
> 1       1       1234567 SMITH-KONES                             AMY
> 1       1       1234567 Garrett                                 Thomas
> V
> 69      1       1234567 O NEAL                                  COLLEEN
> 1       1       1234567 O NEAL                                  COLLEEN
> 5       1       1234567 FULLER                                  ALI
> S
> 2       1       1234567 CALHOUN                                 SLY
> M
> 70      1       1234567 CALHOUN                                 SLY
> M
> 16      1       1234567 DANIEL JR.                              MICHAEL
> J
> 1       1       1234567 DANIEL JR.                              MICHAEL
> J
> 1       1       1234567 DANIEL JR.                              MICHAEL
> J
> 204     1       1234567 KURCZEWSKI-LYONS               DAWN
> M
>
> Here’s some info about the fields shown in the sample data set.
>
> cm_days_in_service = numeric 10 spaces
> cm_valid_ssn_flag = numeric 1 space
> customer_id = numeric 10 spaces
> last_name = character 30 spaces
> first_name = character 20 spaces
> middle_initial = character 1 space
>
> Thanks!
>
> Angie

 | 
Pages: 1
Prev: Lookup table from excel spreadsheet
Next: PROC MIXED/GLIMMIX disagreement