Merge pull request #126 from sasjs/applyformats

Applyformats

all.sas

@@ -484,9 +484,10 @@ https://github.com/yabwon/SAS_PACKAGES/blob/main/packages/baseplus.md#functionex
   contributors of Chris Hemedingers blog [post](
   http://blogs.sas.com/content/sasdummy/2013/06/04/find-a-sas-library-engine/)
 
-  @param libref Library reference (also accepts a 2 level libds ref).
+  @param [in] libref Library reference (also accepts a 2 level libds ref).
 
-  @return output returns the library engine for the FIRST library encountered.
+  @return output returns the library engine (uppercase) for the FIRST library
+    encountered.
 
   @warning will only return the FIRST library engine - for concatenated
     libraries, with different engines, inconsistent results may be encountered.
@@ -518,7 +519,7 @@ https://github.com/yabwon/SAS_PACKAGES/blob/main/packages/baseplus.md#functionex
     %let rc= %sysfunc(close(&dsid));
   %end;
 
-  &engine
+  %upcase(&engine)
 
 %mend mf_getengine;
 
@@ -1532,6 +1533,45 @@ Usage:
   %else %do;1%end;
 
 %mend mf_isint;/**
+  @file
+  @brief Checks whether a string follows correct library.dataset format
+  @details Many macros in the core library accept a library.dataset parameter
+  referred to as 'libds'.  This macro validates the structure of that parameter,
+  eg:
+
+    @li 8 character libref?
+    @li 32 character dataset?
+    @li contains a period?
+
+  It does NOT check whether the dataset exists, or if the library is assigned.
+
+  Usage:
+
+      %put %mf_islibds(work.something)=1;
+      %put %mf_islibds(nolib)=0;
+      %put %mf_islibds(badlibref.ds)=0;
+      %put %mf_islibds(w.t.f)=0;
+
+  @param [in] libds The string to be checked
+
+  @return output Returns 1 if libds is valid, 0 if it is not
+
+  <h4> Related Macros </h4>
+  @li mf_islibds.test.sas
+  @li mp_validatecol.sas
+
+  @version 9.2
+**/
+
+%macro mf_islibds(libds
+)/*/STORE SOURCE*/;
+
+%local regex;
+%let regex=%sysfunc(prxparse(%str(/^[_a-z]\w{0,7}\.[_a-z]\w{0,31}$/i)));
+
+%sysfunc(prxmatch(&regex,&libds))
+
+%mend mf_islibds;/**
   @file
   @brief Returns physical location of various SAS items
   @details Returns location of the PlatformObjectFramework tools
@@ -2225,6 +2265,186 @@ Usage:
 %end;
 
 %mend mp_appendfile;/**
+  @file
+  @brief Apply a set of formats to a table
+  @details Applies a set of formats to one or more SAS datasets.  Can be used
+    to migrate formats from one table to another.  The input table must contain
+    the following columns:
+
+    @li lib - the libref of the table to be updated
+    @li ds - the dataset to be updated
+    @li var - the variable to be updated
+    @li fmt - the format to apply.  Missing or default ($CHAR, 8.) formats are
+      ignored.
+
+  The macro will abort in the following scenarios:
+
+    @li Libref not assigned
+    @li Dataset does not exist
+    @li Input table contains null or invalid values
+
+  Example usage:
+
+      data work.example;
+        set sashelp.prdsale;
+        format _all_ clear;
+      run;
+
+      %mp_getcols(sashelp.prdsale,outds=work.cols)
+
+      data work.cols2;
+        set work.cols;
+        lib='WORK';
+        ds='EXAMPLE';
+        var=name;
+        fmt=format;
+        keep lib ds var fmt;
+      run;
+
+      %mp_applyformats(work.cols2)
+
+  @param [in] inds The input dataset containing the formats to apply (and where
+    to apply them).  Example structure:
+  |LIB:$8.|DS:$32.|VAR:$32.|FMT:$49.|
+  |---|---|---|---|
+  |`WORK `|`EXAMPLE `|`ACTUAL `|`DOLLAR12.2 `|
+  |`WORK `|`EXAMPLE `|`COUNTRY `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`DIVISION `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`MONTH `|`MONNAME3. `|
+  |`WORK `|`EXAMPLE `|`PREDICT `|`DOLLAR12.2 `|
+  |`WORK `|`EXAMPLE `|`PRODTYPE `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`PRODUCT `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`QUARTER `|`8. `|
+  |`WORK `|`EXAMPLE `|`REGION `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`YEAR `|`8. `|
+
+  @param [out] errds= (0) Provide a libds reference here to export the
+    error messages to a table.  In this case, they will not be printed to the
+    log.
+
+  <h4> SAS Macros </h4>
+  @li mf_getengine.sas
+  @li mf_getuniquefileref.sas
+  @li mf_getuniquename.sas
+  @li mf_nobs.sas
+  @li mp_validatecol.sas
+
+
+  <h4> Related Macros </h4>
+  @li mp_getformats.sas
+
+  @version 9.2
+  @author Allan Bowe
+
+**/
+
+%macro mp_applyformats(inds,errds=0
+)/*/STORE SOURCE*/;
+%local outds liblist i engine lib msg ;
+
+/**
+  * Validations
+  */
+proc sort data=&inds;
+  by lib ds var fmt;
+run;
+
+%if &errds=0 %then %let outds=%mf_getuniquename(prefix=mp_applyformats);
+%else %let outds=&errds;
+
+data &outds;
+  set &inds;
+  where fmt not in ('','.', '$', '$CHAR.','8.');
+  length msg $128;
+  by lib ds var fmt;
+  if libref(lib) ne 0 then do;
+    msg=catx(' ','libref',lib,'is not assigned!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+  if exist(cats(lib,'.',ds)) ne 1 then do;
+    msg=catx(' ','libds',lib,'.',ds,'does not exist!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+  %mp_validatecol(fmt,FORMAT,is_fmt)
+  if is_fmt=0 then do;
+    msg=catx(' ','format',fmt,'on libds',lib,'.',ds,'.',var,'is not valid!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+
+  if first.ds then do;
+    retain dsid;
+    dsid=open(cats(lib,'.',ds));
+    if dsid=0 then do;
+      msg=catx(' ','libds',lib,'.',ds,' could not be opened!');
+      %if &errds=0 %then %do;
+        putlog "%str(ERR)OR: " msg;
+      %end;
+      output;
+      return;
+    end;
+    if varnum(dsid,var)<1 then do;
+      msg=catx(' ','Variable',lib,'.',ds,'.',var,' was not found!');
+      %if &errds=0 %then %do;
+        putlog "%str(ERR)OR: " msg;
+      %end;
+      output;
+    end;
+  end;
+  if last.ds then rc=close(dsid);
+run;
+
+proc sql noprint;
+select distinct lib into: liblist separated by ' ' from &inds;
+%put &=liblist;
+%do i=1 %to %sysfunc(countw(&liblist));
+  %let lib=%scan(&liblist,1);
+  %let engine=%mf_getengine(&lib);
+  %if &engine ne V9 and &engine ne BASE %then %do;
+    %let msg=&lib has &engine engine - formats cannot be applied;
+    proc sql;
+    insert into &outds set lib="&lib",ds="_all_",var="_all", msg="&msg" ;
+    %if &errds=0 %then %put %str(ERR)OR: &msg;
+  %end;
+%end;
+
+%if %mf_nobs(&outds)>0 %then %return;
+
+/**
+  * Validations complete - now apply the actual formats!
+  */
+%let fref=%mf_getuniquefileref();
+data _null_;
+  set &inds;
+  by lib ds var fmt;
+  where fmt not in ('','.', '$', '$CHAR.','8.');
+  file &fref;
+  if first.lib then put 'proc datasets nolist lib=' lib ';';
+  if first.ds then put '  modify ' ds ';';
+  put '    format ' var fmt ';';
+  if last.ds then put '  run;';
+  if last.lib then put 'quit;';
+run;
+
+%inc &fref/source2;
+
+%if &errds=0 %then %do;
+  proc sql;
+  drop table &outds;
+%end;
+
+%mend mp_applyformats;/**
   @file
   @brief Generic assertion
   @details Useful in the context of writing sasjs tests.  The results of the
@@ -5855,6 +6075,7 @@ https://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#
 
 
   <h4> Related Macros </h4>
+  @li mp_applyformats.sas
   @li mp_getformats.test.sas
 
   @version 9.2
@@ -7640,8 +7861,6 @@ lock &libds clear;
     according to the variable types and formats.
 
     TODO:
-      @li Respect PKs
-      @li Respect NOT NULLs
       @li Consider dates, datetimes, times, integers etc
 
   Usage:
@@ -7657,16 +7876,23 @@ lock &libds clear;
       );
       %mp_makedata(work.example)
 
-  @param [in] libds The empty table in which to create data
-  @param [out] obs= (500) The number of records to create.
+  @param [in] libds The empty table (libref.dataset) in which to create data
+  @param [out] obs= (500) The maximum number of records to create.  The table
+    is sorted with nodup on the primary key, so the actual number of records may
+    be lower than this.
 
   <h4> SAS Macros </h4>
   @li mf_getuniquename.sas
   @li mf_getvarlen.sas
+  @li mf_getvarlist.sas
+  @li mf_islibds.sas
   @li mf_nobs.sas
   @li mp_getcols.sas
   @li mp_getpk.sas
 
+  <h4> Related Macros </h4>
+  @li mp_makedata.test.sas
+
   @version 9.2
   @author Allan Bowe
 
@@ -7674,45 +7900,59 @@ lock &libds clear;
 
 %macro mp_makedata(libds
   ,obs=500
+  ,seed=1
 )/*/STORE SOURCE*/;
 
-%local ds1 c1 n1 i col charvars numvars;
+%local ds1 ds2 lib ds pk_fields i col charvars numvars ispk;
 
-%if %mf_nobs(&libds)>0 %then %do;
+%if %mf_islibds(&libds)=0 %then %do;
+  %put &sysmacroname: Invalid libds (&libds) - should be library.dataset format;
+  %return;
+%end;
+%else %if %mf_nobs(&libds)>0 %then %do;
   %put &sysmacroname: &libds has data, it will not be recreated;
   %return;
 %end;
 
-%local ds1 c1 n1;
-%let ds1=%mf_getuniquename(prefix=mp_makedata);
-%let c1=%mf_getuniquename(prefix=mp_makedatacol);
-%let n1=%mf_getuniquename(prefix=mp_makedatacol);
-data &ds1;
+/* set up temporary vars */
+%let ds1=%mf_getuniquename(prefix=mp_makedatads1);
+%let ds2=%mf_getuniquename(prefix=mp_makedatads2);
+%let lib=%scan(&libds,1,.);
+%let ds=%scan(&libds,2,.);
+
+/* grab the primary key vars */
+%mp_getpk(&lib,ds=&ds,outds=&ds1)
+
+proc sql noprint;
+select coalescec(pk_fields,'_all_') into: pk_fields from &ds1;
+
+data &ds2;
   if 0 then set &libds;
   do _n_=1 to &obs;
-    &c1=repeat(uuidgen(),10);
-    &n1=ranuni(1)*5000000;
-    drop &c1 &n1;
     %let charvars=%mf_getvarlist(&libds,typefilter=C);
     %if &charvars ^= %then %do i=1 %to %sysfunc(countw(&charvars));
       %let col=%scan(&charvars,&i);
-      &col=subpad(&c1,1,%mf_getvarlen(&libds,&col));
+      /* create random value based on observation number and colum length */
+      &col=repeat(put(md5(cats(_n_)),$hex32.),%mf_getvarlen(&libds,&col)/32);
     %end;
 
     %let numvars=%mf_getvarlist(&libds,typefilter=N);
     %if &numvars ^= %then %do i=1 %to %sysfunc(countw(&numvars));
       %let col=%scan(&numvars,&i);
-      &col=&n1;
+      &col=_n_;
     %end;
     output;
   end;
 run;
+proc sort data=&ds2 nodupkey;
+  by &pk_fields;
+run;
 
-proc append base=&libds data=&ds1;
+proc append base=&libds data=&ds2;
 run;
 
 proc sql;
-drop table &ds1;
+drop table &ds1, &ds2;
 
 %mend mp_makedata;/**
   @file
@@ -9581,10 +9821,13 @@ alter table &libds modify &var char(&len);
       ;;;;
       run;
 
+  Tip - when contributing, use https://regex101.com to test the regex validity!
+
   @param [in] incol The column to be validated
   @param [in] rule The rule to apply.  Current rules:
     @li ISNUM - checks if the variable is numeric
     @li LIBDS - matches LIBREF.DATASET format
+    @li FORMAT - checks if the provided format is syntactically valid
   @param [out] outcol The variable to create, with the results of the match
 
   <h4> SAS Macros </h4>
@@ -9623,6 +9866,20 @@ alter table &libds modify &var char(&len);
   if prxmatch(&tempcol, trim(&incol)) then &outcol=1;
   else &outcol=0;
 %end;
+%else %if &rule=FORMAT %then %do;
+  /* match valid format - regex could probably be improved */
+  if _n_=1 then do;
+    retain &tempcol;
+    &tempcol=prxparse('/^[_a-z\$]\w{0,31}\.[0-9]*$/i');
+    if missing(&tempcol) then do;
+      putlog "%str(ERR)OR: Invalid expression for FORMAT";
+      stop;
+    end;
+    drop &tempcol;
+  end;
+  if prxmatch(&tempcol, trim(&incol)) then &outcol=1;
+  else &outcol=0;
+%end;
 
 %mend mp_validatecol;
 /**

base/mf_getengine.sas

@@ -12,9 +12,10 @@
   contributors of Chris Hemedingers blog [post](
   http://blogs.sas.com/content/sasdummy/2013/06/04/find-a-sas-library-engine/)
 
-  @param libref Library reference (also accepts a 2 level libds ref).
+  @param [in] libref Library reference (also accepts a 2 level libds ref).
 
-  @return output returns the library engine for the FIRST library encountered.
+  @return output returns the library engine (uppercase) for the FIRST library
+    encountered.
 
   @warning will only return the FIRST library engine - for concatenated
     libraries, with different engines, inconsistent results may be encountered.
@@ -46,7 +47,7 @@
     %let rc= %sysfunc(close(&dsid));
   %end;
 
-  &engine
+  %upcase(&engine)
 
 %mend mf_getengine;
 

base/mf_islibds.sas

@@ -0,0 +1,40 @@
+/**
+  @file
+  @brief Checks whether a string follows correct library.dataset format
+  @details Many macros in the core library accept a library.dataset parameter
+  referred to as 'libds'.  This macro validates the structure of that parameter,
+  eg:
+
+    @li 8 character libref?
+    @li 32 character dataset?
+    @li contains a period?
+
+  It does NOT check whether the dataset exists, or if the library is assigned.
+
+  Usage:
+
+      %put %mf_islibds(work.something)=1;
+      %put %mf_islibds(nolib)=0;
+      %put %mf_islibds(badlibref.ds)=0;
+      %put %mf_islibds(w.t.f)=0;
+
+  @param [in] libds The string to be checked
+
+  @return output Returns 1 if libds is valid, 0 if it is not
+
+  <h4> Related Macros </h4>
+  @li mf_islibds.test.sas
+  @li mp_validatecol.sas
+
+  @version 9.2
+**/
+
+%macro mf_islibds(libds
+)/*/STORE SOURCE*/;
+
+%local regex;
+%let regex=%sysfunc(prxparse(%str(/^[_a-z]\w{0,7}\.[_a-z]\w{0,31}$/i)));
+
+%sysfunc(prxmatch(&regex,&libds))
+
+%mend mf_islibds;

base/mp_applyformats.sas

@@ -0,0 +1,181 @@
+/**
+  @file
+  @brief Apply a set of formats to a table
+  @details Applies a set of formats to one or more SAS datasets.  Can be used
+    to migrate formats from one table to another.  The input table must contain
+    the following columns:
+
+    @li lib - the libref of the table to be updated
+    @li ds - the dataset to be updated
+    @li var - the variable to be updated
+    @li fmt - the format to apply.  Missing or default ($CHAR, 8.) formats are
+      ignored.
+
+  The macro will abort in the following scenarios:
+
+    @li Libref not assigned
+    @li Dataset does not exist
+    @li Input table contains null or invalid values
+
+  Example usage:
+
+      data work.example;
+        set sashelp.prdsale;
+        format _all_ clear;
+      run;
+
+      %mp_getcols(sashelp.prdsale,outds=work.cols)
+
+      data work.cols2;
+        set work.cols;
+        lib='WORK';
+        ds='EXAMPLE';
+        var=name;
+        fmt=format;
+        keep lib ds var fmt;
+      run;
+
+      %mp_applyformats(work.cols2)
+
+  @param [in] inds The input dataset containing the formats to apply (and where
+    to apply them).  Example structure:
+  |LIB:$8.|DS:$32.|VAR:$32.|FMT:$49.|
+  |---|---|---|---|
+  |`WORK `|`EXAMPLE `|`ACTUAL `|`DOLLAR12.2 `|
+  |`WORK `|`EXAMPLE `|`COUNTRY `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`DIVISION `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`MONTH `|`MONNAME3. `|
+  |`WORK `|`EXAMPLE `|`PREDICT `|`DOLLAR12.2 `|
+  |`WORK `|`EXAMPLE `|`PRODTYPE `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`PRODUCT `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`QUARTER `|`8. `|
+  |`WORK `|`EXAMPLE `|`REGION `|`$CHAR10. `|
+  |`WORK `|`EXAMPLE `|`YEAR `|`8. `|
+
+  @param [out] errds= (0) Provide a libds reference here to export the
+    error messages to a table.  In this case, they will not be printed to the
+    log.
+
+  <h4> SAS Macros </h4>
+  @li mf_getengine.sas
+  @li mf_getuniquefileref.sas
+  @li mf_getuniquename.sas
+  @li mf_nobs.sas
+  @li mp_validatecol.sas
+
+
+  <h4> Related Macros </h4>
+  @li mp_getformats.sas
+
+  @version 9.2
+  @author Allan Bowe
+
+**/
+
+%macro mp_applyformats(inds,errds=0
+)/*/STORE SOURCE*/;
+%local outds liblist i engine lib msg ;
+
+/**
+  * Validations
+  */
+proc sort data=&inds;
+  by lib ds var fmt;
+run;
+
+%if &errds=0 %then %let outds=%mf_getuniquename(prefix=mp_applyformats);
+%else %let outds=&errds;
+
+data &outds;
+  set &inds;
+  where fmt not in ('','.', '$', '$CHAR.','8.');
+  length msg $128;
+  by lib ds var fmt;
+  if libref(lib) ne 0 then do;
+    msg=catx(' ','libref',lib,'is not assigned!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+  if exist(cats(lib,'.',ds)) ne 1 then do;
+    msg=catx(' ','libds',lib,'.',ds,'does not exist!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+  %mp_validatecol(fmt,FORMAT,is_fmt)
+  if is_fmt=0 then do;
+    msg=catx(' ','format',fmt,'on libds',lib,'.',ds,'.',var,'is not valid!');
+    %if &errds=0 %then %do;
+      putlog "%str(ERR)OR: " msg;
+    %end;
+    output;
+    return;
+  end;
+
+  if first.ds then do;
+    retain dsid;
+    dsid=open(cats(lib,'.',ds));
+    if dsid=0 then do;
+      msg=catx(' ','libds',lib,'.',ds,' could not be opened!');
+      %if &errds=0 %then %do;
+        putlog "%str(ERR)OR: " msg;
+      %end;
+      output;
+      return;
+    end;
+    if varnum(dsid,var)<1 then do;
+      msg=catx(' ','Variable',lib,'.',ds,'.',var,' was not found!');
+      %if &errds=0 %then %do;
+        putlog "%str(ERR)OR: " msg;
+      %end;
+      output;
+    end;
+  end;
+  if last.ds then rc=close(dsid);
+run;
+
+proc sql noprint;
+select distinct lib into: liblist separated by ' ' from &inds;
+%put &=liblist;
+%do i=1 %to %sysfunc(countw(&liblist));
+  %let lib=%scan(&liblist,1);
+  %let engine=%mf_getengine(&lib);
+  %if &engine ne V9 and &engine ne BASE %then %do;
+    %let msg=&lib has &engine engine - formats cannot be applied;
+    proc sql;
+    insert into &outds set lib="&lib",ds="_all_",var="_all", msg="&msg" ;
+    %if &errds=0 %then %put %str(ERR)OR: &msg;
+  %end;
+%end;
+
+%if %mf_nobs(&outds)>0 %then %return;
+
+/**
+  * Validations complete - now apply the actual formats!
+  */
+%let fref=%mf_getuniquefileref();
+data _null_;
+  set &inds;
+  by lib ds var fmt;
+  where fmt not in ('','.', '$', '$CHAR.','8.');
+  file &fref;
+  if first.lib then put 'proc datasets nolist lib=' lib ';';
+  if first.ds then put '  modify ' ds ';';
+  put '    format ' var fmt ';';
+  if last.ds then put '  run;';
+  if last.lib then put 'quit;';
+run;
+
+%inc &fref/source2;
+
+%if &errds=0 %then %do;
+  proc sql;
+  drop table &outds;
+%end;
+
+%mend mp_applyformats;

base/mp_getformats.sas

@@ -48,6 +48,7 @@ https://support.sas.com/documentation/cdl/en/proc/61895/HTML/default/viewer.htm#
 
 
   <h4> Related Macros </h4>
+  @li mp_applyformats.sas
   @li mp_getformats.test.sas
 
   @version 9.2

base/mp_makedata.sas

@@ -10,8 +10,6 @@
     according to the variable types and formats.
 
     TODO:
-      @li Respect PKs
-      @li Respect NOT NULLs
       @li Consider dates, datetimes, times, integers etc
 
   Usage:
@@ -27,16 +25,23 @@
       );
       %mp_makedata(work.example)
 
-  @param [in] libds The empty table in which to create data
-  @param [out] obs= (500) The number of records to create.
+  @param [in] libds The empty table (libref.dataset) in which to create data
+  @param [out] obs= (500) The maximum number of records to create.  The table
+    is sorted with nodup on the primary key, so the actual number of records may
+    be lower than this.
 
   <h4> SAS Macros </h4>
   @li mf_getuniquename.sas
   @li mf_getvarlen.sas
+  @li mf_getvarlist.sas
+  @li mf_islibds.sas
   @li mf_nobs.sas
   @li mp_getcols.sas
   @li mp_getpk.sas
 
+  <h4> Related Macros </h4>
+  @li mp_makedata.test.sas
+
   @version 9.2
   @author Allan Bowe
 
@@ -44,44 +49,58 @@
 
 %macro mp_makedata(libds
   ,obs=500
+  ,seed=1
 )/*/STORE SOURCE*/;
 
-%local ds1 c1 n1 i col charvars numvars;
+%local ds1 ds2 lib ds pk_fields i col charvars numvars ispk;
 
-%if %mf_nobs(&libds)>0 %then %do;
+%if %mf_islibds(&libds)=0 %then %do;
+  %put &sysmacroname: Invalid libds (&libds) - should be library.dataset format;
+  %return;
+%end;
+%else %if %mf_nobs(&libds)>0 %then %do;
   %put &sysmacroname: &libds has data, it will not be recreated;
   %return;
 %end;
 
-%local ds1 c1 n1;
-%let ds1=%mf_getuniquename(prefix=mp_makedata);
-%let c1=%mf_getuniquename(prefix=mp_makedatacol);
-%let n1=%mf_getuniquename(prefix=mp_makedatacol);
-data &ds1;
+/* set up temporary vars */
+%let ds1=%mf_getuniquename(prefix=mp_makedatads1);
+%let ds2=%mf_getuniquename(prefix=mp_makedatads2);
+%let lib=%scan(&libds,1,.);
+%let ds=%scan(&libds,2,.);
+
+/* grab the primary key vars */
+%mp_getpk(&lib,ds=&ds,outds=&ds1)
+
+proc sql noprint;
+select coalescec(pk_fields,'_all_') into: pk_fields from &ds1;
+
+data &ds2;
   if 0 then set &libds;
   do _n_=1 to &obs;
-    &c1=repeat(uuidgen(),10);
-    &n1=ranuni(1)*5000000;
-    drop &c1 &n1;
     %let charvars=%mf_getvarlist(&libds,typefilter=C);
     %if &charvars ^= %then %do i=1 %to %sysfunc(countw(&charvars));
       %let col=%scan(&charvars,&i);
-      &col=subpad(&c1,1,%mf_getvarlen(&libds,&col));
+      /* create random value based on observation number and colum length */
+      &col=repeat(put(md5(cats(_n_)),$hex32.),%mf_getvarlen(&libds,&col)/32);
     %end;
 
     %let numvars=%mf_getvarlist(&libds,typefilter=N);
     %if &numvars ^= %then %do i=1 %to %sysfunc(countw(&numvars));
       %let col=%scan(&numvars,&i);
-      &col=&n1;
+      &col=_n_;
     %end;
     output;
   end;
 run;
+proc sort data=&ds2 nodupkey;
+  by &pk_fields;
+run;
 
-proc append base=&libds data=&ds1;
+proc append base=&libds data=&ds2;
 run;
 
 proc sql;
-drop table &ds1;
+drop table &ds1, &ds2;
 
 %mend mp_makedata;

base/mp_validatecol.sas

@@ -20,10 +20,13 @@
       ;;;;
       run;
 
+  Tip - when contributing, use https://regex101.com to test the regex validity!
+
   @param [in] incol The column to be validated
   @param [in] rule The rule to apply.  Current rules:
     @li ISNUM - checks if the variable is numeric
     @li LIBDS - matches LIBREF.DATASET format
+    @li FORMAT - checks if the provided format is syntactically valid
   @param [out] outcol The variable to create, with the results of the match
 
   <h4> SAS Macros </h4>
@@ -62,5 +65,19 @@
   if prxmatch(&tempcol, trim(&incol)) then &outcol=1;
   else &outcol=0;
 %end;
+%else %if &rule=FORMAT %then %do;
+  /* match valid format - regex could probably be improved */
+  if _n_=1 then do;
+    retain &tempcol;
+    &tempcol=prxparse('/^[_a-z\$]\w{0,31}\.[0-9]*$/i');
+    if missing(&tempcol) then do;
+      putlog "%str(ERR)OR: Invalid expression for FORMAT";
+      stop;
+    end;
+    drop &tempcol;
+  end;
+  if prxmatch(&tempcol, trim(&incol)) then &outcol=1;
+  else &outcol=0;
+%end;
 
 %mend mp_validatecol;

tests/crossplatform/mf_islibds.test.sas

@@ -0,0 +1,46 @@
+/**
+  @file
+  @brief Testing mf_islibds macro
+
+      %put %mf_islibds(work.something)=1;
+      %put %mf_islibds(nolib)=0;
+      %put %mf_islibds(badlibref.ds)=0;
+      %put %mf_islibds(w.t.f)=0;
+
+  <h4> SAS Macros </h4>
+  @li mf_islibds.sas
+  @li mp_assert.sas
+
+**/
+
+%mp_assert(
+  iftrue=(
+    %mf_islibds(work.something)=1
+  ),
+  desc=%str(Checking mf_islibds(work.something)=1),
+  outds=work.test_results
+)
+
+%mp_assert(
+  iftrue=(
+    %mf_islibds(nolib)=0
+  ),
+  desc=%str(Checking mf_islibds(nolib)=0),
+  outds=work.test_results
+)
+
+%mp_assert(
+  iftrue=(
+    %mf_islibds(badlibref.ds)=0
+  ),
+  desc=%str(Checking mf_islibds(badlibref.ds)=0),
+  outds=work.test_results
+)
+
+%mp_assert(
+  iftrue=(
+    %mf_islibds(w.t.f)=0
+  ),
+  desc=%str(Checking mf_islibds(w.t.f)=0),
+  outds=work.test_results
+)

tests/crossplatform/mp_applyformats.test.sas

@@ -0,0 +1,45 @@
+/**
+  @file
+  @brief Testing mp_applyformats.sas macro
+
+  <h4> SAS Macros </h4>
+  @li mf_getvarformat.sas
+  @li mp_applyformats.sas
+  @li mp_assert.sas
+  @li mp_getcols.sas
+
+**/
+
+/**
+  * Test 1 Base case
+  */
+
+data work.example;
+  set sashelp.prdsale;
+  format _all_;
+run;
+%let origfmt=%mf_getvarformat(work.example,month);
+
+%mp_getcols(sashelp.prdsale,outds=work.cols)
+
+data work.cols2;
+  set work.cols;
+  lib='WORK';
+  ds='EXAMPLE';
+  var=name;
+  fmt=format;
+  keep lib ds var fmt;
+run;
+
+%mp_applyformats(work.cols2)
+
+%mp_assert(
+  iftrue=("&orig_fmt"=""),
+  desc=Check that formats were cleared,
+  outds=work.test_results
+)
+%mp_assert(
+  iftrue=("%mf_getvarformat(work.example,month)"="MONNAME3."),
+  desc=Check that formats were applied,
+  outds=work.test_results
+)

tests/crossplatform/mp_makedata.test.sas

@@ -0,0 +1,59 @@
+/**
+  @file
+  @brief Testing mp_makedata.sas macro
+
+  <h4> SAS Macros </h4>
+  @li mf_nobs.sas
+  @li mp_makedata.sas
+  @li mp_assert.sas
+
+**/
+
+/**
+  * Test 1 - Regular makedata call
+  */
+
+proc sql;
+create table work.example(
+  TX_FROM float format=datetime19.,
+  DD_TYPE char(16),
+  DD_SOURCE char(2048),
+  DD_SHORTDESC char(256),
+  constraint pk primary key(tx_from, dd_type,dd_source),
+  constraint nnn not null(DD_SHORTDESC)
+);
+%mp_makedata(work.example,obs=500)
+
+%mp_assert(
+  iftrue=("%mf_nobs(work.example)"="500"),
+  desc=Check that 500 rows were created,
+  outds=work.test_results
+)
+
+data _null_;
+  set work.example;
+  call symputx('lenvar',length(dd_source));
+  stop;
+run;
+%mp_assert(
+  iftrue=("&lenvar"="2048"),
+  desc=Check that entire length of variable is populated,
+  outds=work.test_results
+)
+
+
+proc sql;
+create table work.example2(
+  TX_FROM float format=datetime19.,
+  DD_TYPE char(16),
+  DD_SOURCE char(2048),
+  DD_SHORTDESC char(256),
+  some_num num
+);
+%mp_makedata(work.example2)
+
+%mp_assert(
+  iftrue=(&syscc=0),
+  desc=Ensure tables without keys still generate,
+  outds=work.test_results
+)

tests/crossplatform/mp_validatecol.test.sas

@@ -59,4 +59,37 @@ run;
   desc=Test2 - ISNUM,
   test=EQUALS 4,
   outds=work.test_results
+)
+
+/**
+  * Test 3 - FORMAT
+  */
+data test3;
+  infile datalines4 dsd;
+  input;
+  infile=_infile_;
+  %mp_validatecol(infile,FORMAT,is_format)
+  if is_format=1;
+datalines4;
+$.
+$format.
+$format12.2
+somenum.
+somenum12.4
+above are good
+the rest are bad
+%abort
+1&somethingverybad.
+&
++-1
+.
+a.A
+$format12.1b
+$format12.1b1
+;;;;
+run;
+%mp_assertdsobs(work.test3,
+  desc=Test3 - ISFORMAT,
+  test=EQUALS 5,
+  outds=work.test_results
 )