try in a data step (or SQL equivalent) timestat = intnx('MINUTE2', timestamp, 0, 'END') ; Richard Message was edited by: Richard Carson - added missing parameter ... View more

Try weekstart_dt=intnx('WEEK', '01JAN2014'D, weeknum) ; weekend_dt=intnx('WEEK', '01JAN2014'D, weeknum, 'END') ; Richard ... View more

Another option is to use the TYPES statement which allows specification of subtotals by required combinations of class variables without having to specify values of _TYPE_. Base SAS(R) 9.2 Procedures Guide Richard ... View more

If your table is very large, but is either all numeric or all character, you can modify the above code to apply only to the first row of the table Proc transpose data = have (obs = 1)                          out = pivot (drop = _LABEL_)                ; etc. Then use the SQL Union set operator to append the rest of the data.  Union does not check that the variable names are the same Proc SQL ;      create table want2 as           select * from want                union           select * from have      ; Quit. If you have a mix of character and numeric and you want to span them all with var1 - var300 labelling (why on earth?) then you will need to specify either lengths or formats for the renamed variables before using the sql union to append the data: Data want2 ;      Length var1 8                     var2 $20                     var3 $1                     var4 - var5 8                     ; Run ; Proc SQL ;  /*as above */ Richard ... View more

I'll admit this is a bit of a kludge but a double transposition would probably work, although it might be impractical for a very large dataset (> 10K rows), assuming all your variables are all numeric (or all character).  You probably will want to transpose by id and leave that un-renamed but that's up to you.  Leave out the first drop option if your variables do not have labels. Proc transpose data = have                          out = pivot (drop = _LABEL_)                ;      Var     _ALL_ ;      *     by id ;     /*optional*/ Run ; Proc transpose data = pivot (rename = (_NAME_ = _LABEL_))                          out = want (drop = _NAME_)                          prefix = var                          ;      *     by id ;     /*use it if you used it first time*/ Run ; The general idea is that the second transpose will use the prefix to construct column names because the _NAME_ column no longer exists - instead it will supply labels to the output dataset. BTW this is untested code... Richard ... View more

I think the easiest way is to copy the file to a temporary location, say C:\temp\90mfips.txt.   To preserve the metadata in the indentation of the name or title I suggest creating 'Level' flag, and an accompanying format that specifies the type of name: Proc Format ;   Value $Name_Title   '1' = 'MSA/CMSA title'   '2' = 'PMSA title'   '3' = 'Component county name'   '4' = 'Component city or town name (in New England only)'   ;   Value $Centrl   '1' = 'Central or city'   '2' = 'Outlying or town'   ; You can then strip out all the extra (non data) rows, including the column headers and the layout information at the end of the temporary copy of the file after using the metadata (AKA File Layout) to create column definitions and use column input.  Add 4 extra temporary flags (P1 - P4) to detect the name indentation. The following code is untested so might require some tweaking - Labels might be truncated. Data    Met_Areas ;     Infile    'C:\temp\90mfips.txt' ;     Length    MSA_CMSA    $4         PMSA        $4         Alt_CMSA    $2         State        $2         County        $3         Central_Outlying    $1         Entity        $5         Name_Title    $51         Level        $1         P1 - P4        $1                    ;     Format    Level    $Name_Title.        Central_Outlying  $Centrl.        ;        Label    MSA_CMSA    = 'Four-digit FIPS MSA/CMSA code (6/30/90 definition)'         PMSA        = 'Four-digit FIPS PMSA code (6/30/90 definition)'         Alt_CMSA    = 'Alternative two-digit FIPS CMSA code (6/30/90 definition)'         State        = 'Two-digit FIPS state code (blank at MA level)'         County        = 'Three-digit FIPS county code (blank at MA level)'         Central_Outlying                 = 'Central/Outlying county or city/town flag (1 = central, 2 = Outlying)'         Entity        = 'Five-digit FIPS entity code (blank at MA and county levels)'         Name_Title    = 'Name or Title'         Level        = 'Name or Title Level'         ;        Input    MSA_CMSA    $    1-4         PMSA        $    9-12         Alt_CMSA    $    17-18         State        $    25-26         County        $    27-29         Central_Outlying    $    33         Entity        $    41-45         Name_Title    $    49-99         P1        $    49         P2        $    52         P3        $    57         P4        $    65         ;            Select ;         When (P1 > ' ')    Level = '1' ;         When (P2 > ' ')    Level = '2' ;         When (P3 > ' ')    Level = '3' ;         When (P4 > ' ')    Level = '4' ;         Otherwise ;     End ;     If    MSA_CMSA = ' '    then Delete ;     Name_Title    =    Left (Name_Title) ;     Drop    P1 - P4 ; Run ; Richard ... View more