Merge pull request #81 from sasjs/issue80

fix: removing nonprintables from cards data.  Closes #80

all.sas

@@ -1841,6 +1841,7 @@ Usage:
       put ",""SYSERRORTEXT"" : " syserrortext;
       put ",""SYSHOSTNAME"" : ""&syshostname"" ";
       put ",""SYSJOBID"" : ""&sysjobid"" ";
+      put ",""SYSSCPL"" : ""&sysscpl"" ";
       put ",""SYSSITE"" : ""&syssite"" ";
       sysvlong=quote(trim(symget('sysvlong')));
       put ',"SYSVLONG" : ' sysvlong;
@@ -1855,7 +1856,7 @@ Usage:
 
     %if "&sysprocessmode " = "SAS Stored Process Server " %then %do;
       data _null_;
-        putlog 'stpsrvset program error and syscc';
+        putlog 'stpsrvset program err and syscc';
         rc=stpsrvset('program error', 0);
         call symputx("syscc",0,"g");
       run;
@@ -1901,6 +1902,62 @@ Usage:
 %mend mp_abort;
 
 /** @endcond *//**
+  @file
+  @brief Append (concatenate) two or more files.
+  @details Will append one more more `appendrefs` (filerefs) to a `baseref`.
+  Uses a binary mechanism, so will work with any file type.  For that reason -
+  use with care!   And supply your own trailing carriage returns in each file..
+
+  Usage:
+
+        filename tmp1 temp;
+        filename tmp2 temp;
+        filename tmp3 temp;
+        data _null_; file tmp1; put 'base file';
+        data _null_; file tmp2; put 'append1';
+        data _null_; file tmp3; put 'append2';
+        run;
+        %mp_appendfile(baseref=tmp1, appendrefs=tmp2 tmp3)
+
+
+  @param [in] baseref= Fileref of the base file (should exist)
+  @param [in] appendrefs= One or more filerefs to be appended to the base
+    fileref.  Space separated.
+
+  @version 9.2
+  @author Allan Bowe, source: https://github.com/sasjs/core
+
+  <h4> SAS Macros </h4>
+  @li mp_abort.sas
+  @li mp_binarycopy.sas
+
+
+**/
+
+%macro mp_appendfile(
+  baseref=0,
+  appendrefs=0
+)/*/STORE SOURCE*/;
+
+%mp_abort(iftrue= (&baseref=0)
+  ,mac=&sysmacroname
+  ,msg=%str(Baseref NOT specified!)
+)
+%mp_abort(iftrue= (&appendrefs=0)
+  ,mac=&sysmacroname
+  ,msg=%str(Appendrefs NOT specified!)
+)
+
+%local i;
+%do i=1 %to %sysfunc(countw(&appendrefs));
+  %mp_abort(iftrue= (&syscc>0)
+    ,mac=&sysmacroname
+    ,msg=%str(syscc=&syscc)
+  )
+  %mp_binarycopy(inref=%scan(&appendrefs,&i), outref=&baseref, mode=APPEND)
+%end;
+
+%mend mp_appendfile;/**
   @file
   @brief Generic assertion
   @details Useful in the context of writing sasjs tests.  The results of the
@@ -2131,6 +2188,7 @@ Usage:
 
   <h4> SAS Macros </h4>
   @li mf_existds.sas
+  @li mf_getuniquename.sas
   @li mf_nobs.sas
   @li mp_abort.sas
 
@@ -2216,6 +2274,26 @@ Usage:
   select count(*) into: orig from &lib..&ds;
   quit;
 
+  %local notfound tmp1 tmp2;
+  %let tmp1=%mf_getuniquename();
+  %let tmp2=%mf_getuniquename();
+
+  /* this is a bit convoluted - but using sql outobs=10 throws warnings */
+  proc sql noprint;
+  create view &tmp1 as
+    select distinct &col
+    from &lib..&ds
+    where &col not in (
+      select &ccol from &clib..&cds
+    );
+  data &tmp2;
+    set &tmp1;
+    if _n_>10 then stop;
+  run;
+  proc sql;
+  select distinct &col  into: notfound separated by ' ' from &tmp2;
+
+
   %mp_abort(iftrue= (&syscc ne 0)
     ,mac=&sysmacroname
     ,msg=%str(syscc=&syscc after macro query)
@@ -2226,7 +2304,7 @@ Usage:
     test_description=symget('desc');
     test_result='FAIL';
     test_comments="&sysmacroname: &lib..&ds..&col has &result values "
-      !!"not in &clib..&cds..&ccol ";
+      !!"not in &clib..&cds..&ccol.. First 10 vals:"!!symget('notfound');
   %if &test=ANYVAL %then %do;
     if &result < &orig then test_result='PASS';
   %end;
@@ -3237,12 +3315,21 @@ run;
   @file
   @brief Create a CARDS file from a SAS dataset.
   @details Uses dataset attributes to convert all data into datalines.
-    Running the generated file will rebuild the original dataset.
+    Running the generated file will rebuild the original dataset.  Includes
+    support for large decimals, binary data, PROCESSED_DTTM columns, and
+    alternative encoding.  If the input dataset is empty, the cards file will
+    still be created.
+
+    Additional support to generate a random sample and max rows.
+
   Usage:
 
       %mp_ds2cards(base_ds=sashelp.class
+        , tgt_ds=work.class
         , cards_file= "C:\temp\class.sas"
-        , maxobs=5)
+        , showlog=NO
+        , maxobs=5
+      )
 
   TODO:
     - labelling the dataset
@@ -3253,15 +3340,24 @@ run;
                   that is converted to a cards file.
   @param [in] tgt_ds= Table that the generated cards file would create.
     Optional - if omitted, will be same as BASE_DS.
-  @param [out] cards_file= Location in which to write the (.sas) cards file
-  @param [in] maxobs= to limit output to the first <code>maxobs</code>
-    observations
-  @param [in] showlog= whether to show generated cards file in the SAS log
-    (YES/NO)
-  @param [in] outencoding= provide encoding value for file statement (eg utf-8)
-  @param [in] append= If NO then will rebuild the cards file if it already
+  @param [out] cards_file= ("%sysfunc(pathname(work))/cardgen.sas") Location in
+    which to write the (.sas) cards file
+  @param [in] maxobs= (max) To limit output to the first <code>maxobs</code>
+    observations, enter an integer here.
+  @param [in] random_sample= (NO) Set to YES to generate a random sample of
+    data.  Can be quite slow.
+  @param [in] showlog= (YES) Whether to show generated cards file in the SAS
+    log.  Valid values:
+    @li YES
+    @li NO
+  @param [in] outencoding= Provide encoding value for file statement (eg utf-8)
+  @param [in] append= (NO) If NO then will rebuild the cards file if it already
     exists, otherwise will append to it.  Used by the mp_lib2cards.sas macro.
 
+  <h4> Related Macros </h4>
+  @li mp_lib2cards.sas
+  @li mp_ds2inserts.sas
+  @li mp_mdtablewrite.sas
 
   @version 9.2
   @author Allan Bowe
@@ -3286,15 +3382,15 @@ run;
 %if (&tgt_ds = ) %then %let tgt_ds=&base_ds;
 %if %index(&tgt_ds,.)=0 %then %let tgt_ds=WORK.%scan(&base_ds,2,.);
 %if ("&outencoding" ne "") %then %let outencoding=encoding="&outencoding";
-%if ("&append" = "") %then %let append=;
+%if ("&append" = "" or "&append" = "NO") %then %let append=;
 %else %let append=mod;
 
 /* get varcount */
 %let nvars=0;
 proc sql noprint;
 select count(*) into: nvars from dictionary.columns
-  where libname="%scan(%upcase(&base_ds),1)"
-    and memname="%scan(%upcase(&base_ds),2)";
+  where upcase(libname)="%scan(%upcase(&base_ds),1)"
+    and upcase(memname)="%scan(%upcase(&base_ds),2)";
 %if &nvars=0 %then %do;
   %put %str(WARN)ING: Dataset &base_ds has no variables, will not be converted.;
   %return;
@@ -3350,8 +3446,8 @@ proc sql
 reset outobs=max;
 create table datalines1 as
   select name,type,length,varnum,format,label from dictionary.columns
-  where libname="%upcase(%scan(&base_ds,1))"
-    and memname="%upcase(%scan(&base_ds,2))";
+  where upcase(libname)="%upcase(%scan(&base_ds,1))"
+    and upcase(memname)="%upcase(%scan(&base_ds,2))";
 
 /**
   Due to long decimals cannot use best. format
@@ -3372,7 +3468,18 @@ data datalines_2;
       ,put(',name,',best32.-l)
       ,substrn(put(',name,',bestd32.-l),1
       ,findc(put(',name,',bestd32.-l),"0","TBK")))');
-  else dataline=name;
+  /**
+    * binary data must be converted, to store in text format.  It is identified
+    * by the presence of the $HEX keyword in the format.
+    */
+  else if upcase(format)=:'$HEX' then
+    dataline=cats('put(trim(',name,'),',format,')');
+  /**
+    * There is no easy way to store line breaks in a cards file.
+    * To discuss this, use: https://github.com/sasjs/core/issues/80
+    * Removing all nonprintables with kw (keep writeable)
+    */
+  else dataline=cats('compress(',name,', ,"kw")');
 run;
 
 proc sql noprint;
@@ -3397,7 +3504,8 @@ data _null_;
 
 
 /* Build input statement */
-  if type='char' then type3=':$char.';
+  if upcase(format)=:'$HEX' then type3=':'!!format;
+  else if type='char' then type3=':$char.';
   str2=put(name,$33.)||type3;
 
 
@@ -3419,11 +3527,12 @@ data _null_;
   file &cards_file. &outencoding lrecl=32767 termstr=nl &append;
   length __attrib $32767;
   if _n_=1 then do;
-    put '/*******************************************************************';
-    put " Datalines for %upcase(%scan(&base_ds,2)) dataset ";
-    put " Generated by %nrstr(%%)mp_ds2cards()";
-    put " Available on github.com/sasjs/core";
-    put '********************************************************************/';
+    put '/**';
+    put '  @file';
+    put "  @brief Datalines for %upcase(%scan(&base_ds,2)) dataset";
+    put "  @details Generated by %nrstr(%%)mp_ds2cards()";
+    put "  Available on github.com/sasjs/core";
+    put '**/';
     put "data &tgt_ds &indexes;";
     put "attrib ";
     %do i = 1 %to &nvars;
@@ -3447,11 +3556,11 @@ data _null_;
       put 'run;';
     end;
     else do;
-      put "infile cards dsd delimiter=',';";
+      put "infile cards dsd;";
       put "input ";
       %do i = 1 %to &nvars.;
         %if(%length(&&input_stmt_&i..)) %then
-          put "   &&input_stmt_&i..";
+          put "  &&input_stmt_&i..";
         ;
       %end;
       put ";";
@@ -4230,6 +4339,70 @@ filename &fref1 clear;
 
 %mend mp_filtervalidate;
 /**
+  @file
+  @brief Creates a dataset with column metadata.
+  @details This macro takes the `proc contents` output and "tidies it up" in the
+  following ways:
+
+    @li Blank labels are filled in with column names
+    @li Formats are reconstructed with default values
+    @li Types such as DATE / TIME / DATETIME are inferred from the formats
+
+  Example usage:
+
+      %mp_getcols(sashelp.airline,outds=work.myds)
+
+  @param ds The dataset from which to obtain column metadata
+  @param outds= (work.cols) The output dataset to create. Sample data:
+  |NAME $|LENGTH 8|VARNUM 8|LABEL $|FORMAT $49|TYPE $1 |DDTYPE $|
+  |---|---|---|---|---|---|---|
+  |AIR|8|2|international airline travel (thousands)|8.|N|NUMERIC|
+  |DATE|8|1|DATE|MONYY.|N|DATE|
+  |REGION|3|3|REGION|$3.|C|CHARACTER|
+
+  <h4> Related Macros </h4>
+  @li mf_getvarlist.sas
+  @li mm_getcols.sas
+
+  @version 9.2
+  @author Allan Bowe
+
+**/
+
+%macro mp_getcols(ds, outds=work.cols);
+
+proc contents noprint data=&ds
+  out=_data_ (keep=name type length label varnum format:);
+run;
+data &outds(keep=name type length varnum format label ddtype);
+  set &syslast(rename=(format=format2 type=type2));
+  name=upcase(name);
+  if type2=2 then do;
+    length format $49.;
+    if format2='' then format=cats('$',length,'.');
+    else if formatl=0 then format=cats(format2,'.');
+    else format=cats(format2,formatl,'.');
+    type='C';
+    ddtype='CHARACTER';
+  end;
+  else do;
+    if format2='' then format=cats(length,'.');
+    else if formatl=0 then format=cats(format2,'.');
+    else if formatd=0 then format=cats(format2,formatl,'.');
+    else format=cats(format2,formatl,'.',formatd);
+    type='N';
+    if format=:'DATETIME' then ddtype='DATETIME';
+    else if format=:'DATE' or format=:'DDMMYY' or format=:'MMDDYY'
+      or format=:'YYMMDD' or format=:'E8601DA' or format=:'B8601DA'
+      or format=:'MONYY'
+    then ddtype='DATE';
+    else if format=:'TIME' then ddtype='TIME';
+    else ddtype='NUMERIC';
+  end;
+  if label='' then label=name;
+run;
+
+%mend mp_getcols;/**
   @file mp_getconstraints.sas
   @brief Get constraint details at column level
   @details Useful for capturing constraints before they are dropped / reapplied
@@ -4270,21 +4443,21 @@ filename &fref1 clear;
 /* must use SQL as proc datasets does not support length changes */
 proc sql noprint;
 create table &outds as
-  select a.TABLE_CATALOG as libref
-    ,a.TABLE_NAME
+  select upcase(a.TABLE_CATALOG) as libref
+    ,upcase(a.TABLE_NAME) as TABLE_NAME
     ,a.constraint_type
     ,a.constraint_name
     ,b.column_name
   from dictionary.TABLE_CONSTRAINTS a
   left join dictionary.constraint_column_usage  b
-  on a.TABLE_CATALOG=b.TABLE_CATALOG
-    and a.TABLE_NAME=b.TABLE_NAME
+  on upcase(a.TABLE_CATALOG)=upcase(b.TABLE_CATALOG)
+    and upcase(a.TABLE_NAME)=upcase(b.TABLE_NAME)
     and a.constraint_name=b.constraint_name
-  where a.TABLE_CATALOG="&lib"
-    and b.TABLE_CATALOG="&lib"
+  where upcase(a.TABLE_CATALOG)="&lib"
+    and upcase(b.TABLE_CATALOG)="&lib"
   %if "&ds" ne "" %then %do;
-    and a.TABLE_NAME="&ds"
-    and b.TABLE_NAME="&ds"
+    and upcase(a.TABLE_NAME)="&ds"
+    and upcase(b.TABLE_NAME)="&ds"
   %end;
   ;
 
@@ -4835,7 +5008,7 @@ run;
   proc sql noprint;
   select sysvalue into: schemaactual
     from dictionary.libnames
-    where libname="&libref" and engine='SQLSVR';
+    where upcase(libname)="&libref" and engine='SQLSVR';
   %let schema=%sysfunc(coalescec(&schemaactual,&schema,&libref));
 
   %do x=1 %to %sysfunc(countw(&dsnlist));
@@ -4928,7 +5101,7 @@ run;
   proc sql noprint;
   select sysvalue into: schemaactual
     from dictionary.libnames
-    where libname="&libref" and engine='POSTGRES';
+    where upcase(libname)="&libref" and engine='POSTGRES';
   %let schema=%sysfunc(coalescec(&schemaactual,&schema,&libref));
   data _null_;
     file &fref mod;
@@ -6030,14 +6203,14 @@ select distinct lowcase(memname)
   We take the standard definition one step further by embedding the informat
   in the table header row, like so:
 
-      |var1:$|var2:best.|var3:date9.|
+      |var1:$32|var2:best.|var3:date9.|
       |---|---|---|
       |some text|42|01JAN1960|
       |blah|1|31DEC1999|
 
   Which resolves to:
 
-  |var1:$|var2:best.|var3:date9.|
+  |var1:$32|var2:best.|var3:date9.|
   |---|---|---|
   |some text|42|01JAN1960|
   |blah|1|31DEC1999|
@@ -9986,6 +10159,7 @@ data _null_;
   put '    put ",""SYSERRORTEXT"" : ""&syserrortext"" "; ';
   put '    put ",""SYSHOSTNAME"" : ""&syshostname"" "; ';
   put '    put ",""SYSJOBID"" : ""&sysjobid"" "; ';
+  put '    put ",""SYSSCPL"" : ""&sysscpl"" "; ';
   put '    put ",""SYSSITE"" : ""&syssite"" "; ';
   put '    sysvlong=quote(trim(symget(''sysvlong''))); ';
   put '    put '',"SYSVLONG" : '' sysvlong; ';
@@ -13481,6 +13655,7 @@ run;
     put ",""SYSERRORTEXT"" : ""&syserrortext"" ";
     put ",""SYSHOSTNAME"" : ""&syshostname"" ";
     put ",""SYSJOBID"" : ""&sysjobid"" ";
+    put ",""SYSSCPL"" : ""&sysscpl"" ";
     put ",""SYSSITE"" : ""&syssite"" ";
     sysvlong=quote(trim(symget('sysvlong')));
     put ',"SYSVLONG" : ' sysvlong;
@@ -14952,6 +15127,7 @@ data _null_;
   put '    put ",""SYSCC"" : ""&syscc"" "; ';
   put '    put ",""SYSERRORTEXT"" : ""&syserrortext"" "; ';
   put '    put ",""SYSHOSTNAME"" : ""&syshostname"" "; ';
+  put '    put ",""SYSSCPL"" : ""&sysscpl"" "; ';
   put '    put ",""SYSSITE"" : ""&syssite"" "; ';
   put '    sysvlong=quote(trim(symget(''sysvlong''))); ';
   put '    put '',"SYSVLONG" : '' sysvlong; ';
@@ -18791,6 +18967,7 @@ filename &fref1 clear;
     put ",""SYSCC"" : ""&syscc"" ";
     put ",""SYSERRORTEXT"" : ""&syserrortext"" ";
     put ",""SYSHOSTNAME"" : ""&syshostname"" ";
+    put ",""SYSSCPL"" : ""&sysscpl"" ";
     put ",""SYSSITE"" : ""&syssite"" ";
     sysvlong=quote(trim(symget('sysvlong')));
     put ',"SYSVLONG" : ' sysvlong;

base/mp_abort.sas

@@ -193,6 +193,7 @@
       put ",""SYSERRORTEXT"" : " syserrortext;
       put ",""SYSHOSTNAME"" : ""&syshostname"" ";
       put ",""SYSJOBID"" : ""&sysjobid"" ";
+      put ",""SYSSCPL"" : ""&sysscpl"" ";
       put ",""SYSSITE"" : ""&syssite"" ";
       sysvlong=quote(trim(symget('sysvlong')));
       put ',"SYSVLONG" : ' sysvlong;
@@ -207,7 +208,7 @@
 
     %if "&sysprocessmode " = "SAS Stored Process Server " %then %do;
       data _null_;
-        putlog 'stpsrvset program error and syscc';
+        putlog 'stpsrvset program err and syscc';
         rc=stpsrvset('program error', 0);
         call symputx("syscc",0,"g");
       run;

base/mp_appendfile.sas

@@ -0,0 +1,57 @@
+/**
+  @file
+  @brief Append (concatenate) two or more files.
+  @details Will append one more more `appendrefs` (filerefs) to a `baseref`.
+  Uses a binary mechanism, so will work with any file type.  For that reason -
+  use with care!   And supply your own trailing carriage returns in each file..
+
+  Usage:
+
+        filename tmp1 temp;
+        filename tmp2 temp;
+        filename tmp3 temp;
+        data _null_; file tmp1; put 'base file';
+        data _null_; file tmp2; put 'append1';
+        data _null_; file tmp3; put 'append2';
+        run;
+        %mp_appendfile(baseref=tmp1, appendrefs=tmp2 tmp3)
+
+
+  @param [in] baseref= Fileref of the base file (should exist)
+  @param [in] appendrefs= One or more filerefs to be appended to the base
+    fileref.  Space separated.
+
+  @version 9.2
+  @author Allan Bowe, source: https://github.com/sasjs/core
+
+  <h4> SAS Macros </h4>
+  @li mp_abort.sas
+  @li mp_binarycopy.sas
+
+
+**/
+
+%macro mp_appendfile(
+  baseref=0,
+  appendrefs=0
+)/*/STORE SOURCE*/;
+
+%mp_abort(iftrue= (&baseref=0)
+  ,mac=&sysmacroname
+  ,msg=%str(Baseref NOT specified!)
+)
+%mp_abort(iftrue= (&appendrefs=0)
+  ,mac=&sysmacroname
+  ,msg=%str(Appendrefs NOT specified!)
+)
+
+%local i;
+%do i=1 %to %sysfunc(countw(&appendrefs));
+  %mp_abort(iftrue= (&syscc>0)
+    ,mac=&sysmacroname
+    ,msg=%str(syscc=&syscc)
+  )
+  %mp_binarycopy(inref=%scan(&appendrefs,&i), outref=&baseref, mode=APPEND)
+%end;
+
+%mend mp_appendfile;

base/mp_assertcolvals.sas

@@ -30,6 +30,7 @@
 
   <h4> SAS Macros </h4>
   @li mf_existds.sas
+  @li mf_getuniquename.sas
   @li mf_nobs.sas
   @li mp_abort.sas
 
@@ -115,6 +116,26 @@
   select count(*) into: orig from &lib..&ds;
   quit;
 
+  %local notfound tmp1 tmp2;
+  %let tmp1=%mf_getuniquename();
+  %let tmp2=%mf_getuniquename();
+
+  /* this is a bit convoluted - but using sql outobs=10 throws warnings */
+  proc sql noprint;
+  create view &tmp1 as
+    select distinct &col
+    from &lib..&ds
+    where &col not in (
+      select &ccol from &clib..&cds
+    );
+  data &tmp2;
+    set &tmp1;
+    if _n_>10 then stop;
+  run;
+  proc sql;
+  select distinct &col  into: notfound separated by ' ' from &tmp2;
+
+
   %mp_abort(iftrue= (&syscc ne 0)
     ,mac=&sysmacroname
     ,msg=%str(syscc=&syscc after macro query)
@@ -125,7 +146,7 @@
     test_description=symget('desc');
     test_result='FAIL';
     test_comments="&sysmacroname: &lib..&ds..&col has &result values "
-      !!"not in &clib..&cds..&ccol ";
+      !!"not in &clib..&cds..&ccol.. First 10 vals:"!!symget('notfound');
   %if &test=ANYVAL %then %do;
     if &result < &orig then test_result='PASS';
   %end;

base/mp_ds2cards.sas

@@ -2,12 +2,21 @@
   @file
   @brief Create a CARDS file from a SAS dataset.
   @details Uses dataset attributes to convert all data into datalines.
-    Running the generated file will rebuild the original dataset.
+    Running the generated file will rebuild the original dataset.  Includes
+    support for large decimals, binary data, PROCESSED_DTTM columns, and
+    alternative encoding.  If the input dataset is empty, the cards file will
+    still be created.
+
+    Additional support to generate a random sample and max rows.
+
   Usage:
 
       %mp_ds2cards(base_ds=sashelp.class
+        , tgt_ds=work.class
         , cards_file= "C:\temp\class.sas"
-        , maxobs=5)
+        , showlog=NO
+        , maxobs=5
+      )
 
   TODO:
     - labelling the dataset
@@ -18,15 +27,24 @@
                   that is converted to a cards file.
   @param [in] tgt_ds= Table that the generated cards file would create.
     Optional - if omitted, will be same as BASE_DS.
-  @param [out] cards_file= Location in which to write the (.sas) cards file
-  @param [in] maxobs= to limit output to the first <code>maxobs</code>
-    observations
-  @param [in] showlog= whether to show generated cards file in the SAS log
-    (YES/NO)
-  @param [in] outencoding= provide encoding value for file statement (eg utf-8)
-  @param [in] append= If NO then will rebuild the cards file if it already
+  @param [out] cards_file= ("%sysfunc(pathname(work))/cardgen.sas") Location in
+    which to write the (.sas) cards file
+  @param [in] maxobs= (max) To limit output to the first <code>maxobs</code>
+    observations, enter an integer here.
+  @param [in] random_sample= (NO) Set to YES to generate a random sample of
+    data.  Can be quite slow.
+  @param [in] showlog= (YES) Whether to show generated cards file in the SAS
+    log.  Valid values:
+    @li YES
+    @li NO
+  @param [in] outencoding= Provide encoding value for file statement (eg utf-8)
+  @param [in] append= (NO) If NO then will rebuild the cards file if it already
     exists, otherwise will append to it.  Used by the mp_lib2cards.sas macro.
 
+  <h4> Related Macros </h4>
+  @li mp_lib2cards.sas
+  @li mp_ds2inserts.sas
+  @li mp_mdtablewrite.sas
 
   @version 9.2
   @author Allan Bowe
@@ -51,15 +69,15 @@
 %if (&tgt_ds = ) %then %let tgt_ds=&base_ds;
 %if %index(&tgt_ds,.)=0 %then %let tgt_ds=WORK.%scan(&base_ds,2,.);
 %if ("&outencoding" ne "") %then %let outencoding=encoding="&outencoding";
-%if ("&append" = "") %then %let append=;
+%if ("&append" = "" or "&append" = "NO") %then %let append=;
 %else %let append=mod;
 
 /* get varcount */
 %let nvars=0;
 proc sql noprint;
 select count(*) into: nvars from dictionary.columns
-  where libname="%scan(%upcase(&base_ds),1)"
-    and memname="%scan(%upcase(&base_ds),2)";
+  where upcase(libname)="%scan(%upcase(&base_ds),1)"
+    and upcase(memname)="%scan(%upcase(&base_ds),2)";
 %if &nvars=0 %then %do;
   %put %str(WARN)ING: Dataset &base_ds has no variables, will not be converted.;
   %return;
@@ -115,8 +133,8 @@ proc sql
 reset outobs=max;
 create table datalines1 as
   select name,type,length,varnum,format,label from dictionary.columns
-  where libname="%upcase(%scan(&base_ds,1))"
-    and memname="%upcase(%scan(&base_ds,2))";
+  where upcase(libname)="%upcase(%scan(&base_ds,1))"
+    and upcase(memname)="%upcase(%scan(&base_ds,2))";
 
 /**
   Due to long decimals cannot use best. format
@@ -137,7 +155,18 @@ data datalines_2;
       ,put(',name,',best32.-l)
       ,substrn(put(',name,',bestd32.-l),1
       ,findc(put(',name,',bestd32.-l),"0","TBK")))');
-  else dataline=name;
+  /**
+    * binary data must be converted, to store in text format.  It is identified
+    * by the presence of the $HEX keyword in the format.
+    */
+  else if upcase(format)=:'$HEX' then
+    dataline=cats('put(trim(',name,'),',format,')');
+  /**
+    * There is no easy way to store line breaks in a cards file.
+    * To discuss this, use: https://github.com/sasjs/core/issues/80
+    * Removing all nonprintables with kw (keep writeable)
+    */
+  else dataline=cats('compress(',name,', ,"kw")');
 run;
 
 proc sql noprint;
@@ -162,7 +191,8 @@ data _null_;
 
 
 /* Build input statement */
-  if type='char' then type3=':$char.';
+  if upcase(format)=:'$HEX' then type3=':'!!format;
+  else if type='char' then type3=':$char.';
   str2=put(name,$33.)||type3;
 
 
@@ -184,11 +214,12 @@ data _null_;
   file &cards_file. &outencoding lrecl=32767 termstr=nl &append;
   length __attrib $32767;
   if _n_=1 then do;
-    put '/*******************************************************************';
-    put " Datalines for %upcase(%scan(&base_ds,2)) dataset ";
-    put " Generated by %nrstr(%%)mp_ds2cards()";
-    put " Available on github.com/sasjs/core";
-    put '********************************************************************/';
+    put '/**';
+    put '  @file';
+    put "  @brief Datalines for %upcase(%scan(&base_ds,2)) dataset";
+    put "  @details Generated by %nrstr(%%)mp_ds2cards()";
+    put "  Available on github.com/sasjs/core";
+    put '**/';
     put "data &tgt_ds &indexes;";
     put "attrib ";
     %do i = 1 %to &nvars;
@@ -212,11 +243,11 @@ data _null_;
       put 'run;';
     end;
     else do;
-      put "infile cards dsd delimiter=',';";
+      put "infile cards dsd;";
       put "input ";
       %do i = 1 %to &nvars.;
         %if(%length(&&input_stmt_&i..)) %then
-          put "   &&input_stmt_&i..";
+          put "  &&input_stmt_&i..";
         ;
       %end;
       put ";";

base/mp_getcols.sas

@@ -0,0 +1,65 @@
+/**
+  @file
+  @brief Creates a dataset with column metadata.
+  @details This macro takes the `proc contents` output and "tidies it up" in the
+  following ways:
+
+    @li Blank labels are filled in with column names
+    @li Formats are reconstructed with default values
+    @li Types such as DATE / TIME / DATETIME are inferred from the formats
+
+  Example usage:
+
+      %mp_getcols(sashelp.airline,outds=work.myds)
+
+  @param ds The dataset from which to obtain column metadata
+  @param outds= (work.cols) The output dataset to create. Sample data:
+  |NAME $|LENGTH 8|VARNUM 8|LABEL $|FORMAT $49|TYPE $1 |DDTYPE $|
+  |---|---|---|---|---|---|---|
+  |AIR|8|2|international airline travel (thousands)|8.|N|NUMERIC|
+  |DATE|8|1|DATE|MONYY.|N|DATE|
+  |REGION|3|3|REGION|$3.|C|CHARACTER|
+
+  <h4> Related Macros </h4>
+  @li mf_getvarlist.sas
+  @li mm_getcols.sas
+
+  @version 9.2
+  @author Allan Bowe
+
+**/
+
+%macro mp_getcols(ds, outds=work.cols);
+
+proc contents noprint data=&ds
+  out=_data_ (keep=name type length label varnum format:);
+run;
+data &outds(keep=name type length varnum format label ddtype);
+  set &syslast(rename=(format=format2 type=type2));
+  name=upcase(name);
+  if type2=2 then do;
+    length format $49.;
+    if format2='' then format=cats('$',length,'.');
+    else if formatl=0 then format=cats(format2,'.');
+    else format=cats(format2,formatl,'.');
+    type='C';
+    ddtype='CHARACTER';
+  end;
+  else do;
+    if format2='' then format=cats(length,'.');
+    else if formatl=0 then format=cats(format2,'.');
+    else if formatd=0 then format=cats(format2,formatl,'.');
+    else format=cats(format2,formatl,'.',formatd);
+    type='N';
+    if format=:'DATETIME' then ddtype='DATETIME';
+    else if format=:'DATE' or format=:'DDMMYY' or format=:'MMDDYY'
+      or format=:'YYMMDD' or format=:'E8601DA' or format=:'B8601DA'
+      or format=:'MONYY'
+    then ddtype='DATE';
+    else if format=:'TIME' then ddtype='TIME';
+    else ddtype='NUMERIC';
+  end;
+  if label='' then label=name;
+run;
+
+%mend mp_getcols;

base/mp_getconstraints.sas

@@ -39,21 +39,21 @@
 /* must use SQL as proc datasets does not support length changes */
 proc sql noprint;
 create table &outds as
-  select a.TABLE_CATALOG as libref
-    ,a.TABLE_NAME
+  select upcase(a.TABLE_CATALOG) as libref
+    ,upcase(a.TABLE_NAME) as TABLE_NAME
     ,a.constraint_type
     ,a.constraint_name
     ,b.column_name
   from dictionary.TABLE_CONSTRAINTS a
   left join dictionary.constraint_column_usage  b
-  on a.TABLE_CATALOG=b.TABLE_CATALOG
-    and a.TABLE_NAME=b.TABLE_NAME
+  on upcase(a.TABLE_CATALOG)=upcase(b.TABLE_CATALOG)
+    and upcase(a.TABLE_NAME)=upcase(b.TABLE_NAME)
     and a.constraint_name=b.constraint_name
-  where a.TABLE_CATALOG="&lib"
-    and b.TABLE_CATALOG="&lib"
+  where upcase(a.TABLE_CATALOG)="&lib"
+    and upcase(b.TABLE_CATALOG)="&lib"
   %if "&ds" ne "" %then %do;
-    and a.TABLE_NAME="&ds"
-    and b.TABLE_NAME="&ds"
+    and upcase(a.TABLE_NAME)="&ds"
+    and upcase(b.TABLE_NAME)="&ds"
   %end;
   ;
 

base/mp_getddl.sas

@@ -211,7 +211,7 @@ run;
   proc sql noprint;
   select sysvalue into: schemaactual
     from dictionary.libnames
-    where libname="&libref" and engine='SQLSVR';
+    where upcase(libname)="&libref" and engine='SQLSVR';
   %let schema=%sysfunc(coalescec(&schemaactual,&schema,&libref));
 
   %do x=1 %to %sysfunc(countw(&dsnlist));
@@ -304,7 +304,7 @@ run;
   proc sql noprint;
   select sysvalue into: schemaactual
     from dictionary.libnames
-    where libname="&libref" and engine='POSTGRES';
+    where upcase(libname)="&libref" and engine='POSTGRES';
   %let schema=%sysfunc(coalescec(&schemaactual,&schema,&libref));
   data _null_;
     file &fref mod;

base/mp_mdtablewrite.sas

@@ -14,14 +14,14 @@
   We take the standard definition one step further by embedding the informat
   in the table header row, like so:
 
-      |var1:$|var2:best.|var3:date9.|
+      |var1:$32|var2:best.|var3:date9.|
       |---|---|---|
       |some text|42|01JAN1960|
       |blah|1|31DEC1999|
 
   Which resolves to:
 
-  |var1:$|var2:best.|var3:date9.|
+  |var1:$32|var2:best.|var3:date9.|
   |---|---|---|
   |some text|42|01JAN1960|
   |blah|1|31DEC1999|

meta/mm_createwebservice.sas

@@ -385,6 +385,7 @@ data _null_;
   put '    put ",""SYSERRORTEXT"" : ""&syserrortext"" "; ';
   put '    put ",""SYSHOSTNAME"" : ""&syshostname"" "; ';
   put '    put ",""SYSJOBID"" : ""&sysjobid"" "; ';
+  put '    put ",""SYSSCPL"" : ""&sysscpl"" "; ';
   put '    put ",""SYSSITE"" : ""&syssite"" "; ';
   put '    sysvlong=quote(trim(symget(''sysvlong''))); ';
   put '    put '',"SYSVLONG" : '' sysvlong; ';

meta/mm_webout.sas

@@ -155,6 +155,7 @@
     put ",""SYSERRORTEXT"" : ""&syserrortext"" ";
     put ",""SYSHOSTNAME"" : ""&syshostname"" ";
     put ",""SYSJOBID"" : ""&sysjobid"" ";
+    put ",""SYSSCPL"" : ""&sysscpl"" ";
     put ",""SYSSITE"" : ""&syssite"" ";
     sysvlong=quote(trim(symget('sysvlong')));
     put ',"SYSVLONG" : ' sysvlong;

tests/crossplatform/mp_appendfile.test.sas

@@ -0,0 +1,41 @@
+/**
+  @file
+  @brief Testing mp_appendfile.sas macro
+
+  <h4> SAS Macros </h4>
+  @li mp_appendfile.sas
+  @li mp_assert.sas
+
+**/
+
+
+filename tmp1 temp;
+filename tmp2 temp;
+filename tmp3 temp;
+data _null_; file tmp1; put 'base file';
+data _null_; file tmp2; put 'append1';
+data _null_; file tmp3; put 'append2';
+run;
+%mp_appendfile(baseref=tmp1, appendrefs=tmp2 tmp3)
+data _null_;
+  infile tmp1;
+  input;
+  put _infile_;
+  call symputx(cats('check',_n_),_infile_);
+run;
+%global check1 check2 check3;
+%mp_assert(
+  iftrue=("&check1"="base file"),
+  desc=Line 1 of file tmp1 is correct,
+  outds=work.test_results
+)
+%mp_assert(
+  iftrue=("&check2"="append1"),
+  desc=Line 2 of file tmp1 is correct,
+  outds=work.test_results
+)
+%mp_assert(
+  iftrue=("&check3"="append2"),
+  desc=Line 3 of file tmp1 is correct,
+  outds=work.test_results
+)

tests/crossplatform/mp_ds2cards.test.sas

@@ -0,0 +1,60 @@
+/**
+  @file
+  @brief Testing mp_ds2cards.sas macro
+
+  <h4> SAS Macros </h4>
+  @li mp_ds2cards.sas
+  @li mp_assert.sas
+
+**/
+
+/**
+  * test 1 - rebuild an existing dataset
+  * Cars is a great dataset - it contains leading spaces, and formatted numerics
+  */
+
+%mp_ds2cards(base_ds=sashelp.cars
+  , tgt_ds=work.test
+  , cards_file= "%sysfunc(pathname(work))/cars.sas"
+  , showlog=NO
+)
+%inc "%sysfunc(pathname(work))/cars.sas"/source2;
+
+proc compare base=sashelp.cars compare=work.test;
+quit;
+
+%mp_assert(
+  iftrue=(&sysinfo=1),
+  desc=sashelp.cars is identical except for ds label,
+  outds=work.test_results
+)
+
+/**
+  * test 2 - binary data compare
+  */
+data work.binarybase;
+  format bin $hex500. z $hex.;
+  do x=1 to 250;
+    z=byte(x);
+    bin=trim(bin)!!z;
+    output;
+  end;
+run;
+
+%mp_ds2cards(base_ds=work.binarybase
+  , showlog=YES
+  , cards_file="%sysfunc(pathname(work))/c2.sas"
+  , tgt_ds=work.binarycompare
+  , append=
+)
+
+%inc "%sysfunc(pathname(work))/c2.sas"/source2;
+
+proc compare base=work.binarybase compare=work.binarycompare;
+run;
+
+%mp_assert(
+  iftrue=(&sysinfo=0),
+  desc=work.binarybase dataset is identical,
+  outds=work.test_results
+)

tests/crossplatform/mp_getcols.test.sas

@@ -0,0 +1,33 @@
+/**
+  @file
+  @brief Testing mp_getcols macro
+
+  <h4> SAS Macros </h4>
+  @li mp_getcols.sas
+  @li mp_assertcolvals.sas
+  @li mp_assertdsobs.sas
+
+**/
+
+
+/* valid filter */
+%mp_getcols(sashelp.airline,outds=work.info)
+
+
+%mp_assertdsobs(work.info,
+  desc=Has 3 records,
+  test=EQUALS 3,
+  outds=work.test_results
+)
+
+data work.check;
+  length val $10;
+  do val='NUMERIC','DATE','CHARACTER';
+    output;
+  end;
+run;
+%mp_assertcolvals(work.info.ddtype,
+  checkvals=work.check.val,
+  desc=All values have a match,
+  test=ALLVALS
+)

viya/mv_createwebservice.sas

@@ -590,6 +590,7 @@ data _null_;
   put '    put ",""SYSCC"" : ""&syscc"" "; ';
   put '    put ",""SYSERRORTEXT"" : ""&syserrortext"" "; ';
   put '    put ",""SYSHOSTNAME"" : ""&syshostname"" "; ';
+  put '    put ",""SYSSCPL"" : ""&sysscpl"" "; ';
   put '    put ",""SYSSITE"" : ""&syssite"" "; ';
   put '    sysvlong=quote(trim(symget(''sysvlong''))); ';
   put '    put '',"SYSVLONG" : '' sysvlong; ';

viya/mv_webout.sas

@@ -215,6 +215,7 @@
     put ",""SYSCC"" : ""&syscc"" ";
     put ",""SYSERRORTEXT"" : ""&syserrortext"" ";
     put ",""SYSHOSTNAME"" : ""&syshostname"" ";
+    put ",""SYSSCPL"" : ""&sysscpl"" ";
     put ",""SYSSITE"" : ""&syssite"" ";
     sysvlong=quote(trim(symget('sysvlong')));
     put ',"SYSVLONG" : ' sysvlong;