feat: mp_csv2ds macro for importing a CSV using a SAS table to provide a template (eg for lengths / types etc)

all.sas

@@ -1904,6 +1904,113 @@ Usage:
 
 %mend;
 /**
+  @file mp_csv2ds.sas
+  @brief Efficient import of arbitrary CSV using a dataset as template
+  @details Used to import relevant columns from a large CSV using
+  a dataset to provide the types and lengths.  Assumes that a header
+  row is provided, and datarows start on line 2.  Extra columns in
+  both the CSV and base dataset are ignored.
+
+  Usage:
+
+      filename mycsv temp;
+      data _null_;
+        file mycsv;
+        put 'name,age,nickname';
+        put 'John,48,Jonny';
+        put 'Jennifer,23,Jen';
+      run;
+
+      %mp_csv2ds(inref=mycsv,outds=myds,baseds=sashelp.class)
+
+
+  @param inref= fileref to the CSV
+  @param outds= output ds.  Could also be a view (eg `outds=myds/view=myds`)
+  @param baseds= Template dataset on which to create the input statement.
+    Is used to determine types, lengths, and any informats.
+
+  @version 9.2
+  @author Allan Bowe
+**/
+
+%macro mp_csv2ds(inref=0,outds=0,baseds=0);
+%if &inref=0 %then %do;
+  %put %str(ERR)OR: the INREF variable must be provided;
+  %let syscc=4;
+  %abort;
+%end;
+%if &outds=0 %then %do;
+  %put %str(ERR)OR: the OUTDS variable must be provided;
+  %let syscc=4;
+  %return;
+%end;
+%if &baseds=0 %then %do;
+  %put %str(ERR)OR: the BASEDS variable must be provided;
+  %let syscc=4;
+  %return;
+%end;
+
+/* get the variables in the CSV */
+data _data_;
+  infile &inref;
+  input;
+  length name $32;
+  do i=1 to countc(_infile_,',')+1;
+    name=upcase(scan(_infile_,i,','));
+    output;
+  end;
+  stop;
+run;
+%local csv_vars;%let csv_vars=&syslast;
+
+/* get the variables in the dataset */
+proc contents noprint data=&baseds
+  out=_data_ (keep=name type length format: informat);
+run;
+%local base_vars; %let base_vars=&syslast;
+
+proc sql undo_policy=none;
+create table &csv_vars as
+  select a.*
+    ,b.type
+    ,b.length
+    ,b.format
+    ,b.formatd
+    ,b.formatl
+    ,b.informat
+  from &csv_vars a
+  left join &base_vars b
+  on a.name=upcase(b.name)
+  order by i;
+
+/* prepare the input statement */
+%local instat dropvars;
+data _null_;
+  set &syslast end=last;
+  length in dropvars $32767;
+  retain in dropvars;
+  if missing(type) then do;
+    informat='$1.';
+    dropvars=catx(' ',dropvars,name);
+  end;
+  else if missing(informat) then do;
+    if type=1 then informat='best.';
+    else informat=cats('$',length,'.');
+  end;
+  in=catx(' ',in,name,':',informat);
+  if last then do;
+    call symputx('instat',in,'l');
+    call symputx('dropvars',dropvars,'l');
+  end;
+run;
+
+data &outds;
+  infile &inref dsd firstobs=2;
+  input &instat;
+  drop &dropvars;
+run;
+
+%mend;/**
   @file mp_deleteconstraints.sas
   @brief Delete constraionts
   @details Takes the output from mp_getconstraints.sas as input
@@ -2041,7 +2148,8 @@ data &outds (compress=no keep=file_or_folder filepath filename ext msg directory
   dnum = dnum(did);
   do i = 1 to dnum;
     filename = dread(did, i);
-    rc = filename(fref2, cats(directory,'/',filename));
+    filepath=cats(directory,'/',filename);
+    rc = filename(fref2,filepath);
     midd=dopen(fref2);
     dmsg=sysmsg();
     if did > 0 then file_or_folder='folder';
@@ -2064,7 +2172,6 @@ data &outds (compress=no keep=file_or_folder filepath filename ext msg directory
       ext='';
       file_or_folder='folder';
     end;
-    filepath="&path/"!!filename;
     output;
   end;
   rc = dclose(did);
@@ -2642,7 +2749,7 @@ run;
     %let curds=%scan(&dsnlist,&x);
     data _null_;
       file &fref mod;
-      length nm lab $1024;
+      length nm lab $1024 typ $20;
       set &colinfo (where=(upcase(memname)="&curds")) end=last;
 
       if _n_=1 then do;
@@ -2656,10 +2763,12 @@ run;
       end;
       else put "   ,"@@;
       if length(format)>1 then fmt=" format="!!cats(format);
-      len=" length="!!cats(length);
+      if length(label)>1 then lab=" label="!!quote(trim(label));
-      lab=" label="!!quote(trim(label));
       if notnull='yes' then notnul=' not null';
-      put name type len fmt notnul lab;
+      if type='char' then typ=cats('char(',length,')');
+      else if length ne 8 then typ='num length='!!left(length);
+      else typ='num';
+      put name typ fmt notnul lab;
     run;
 
     /* Extra step for data constraints */

base/mf_getengine.sas

@@ -3,7 +3,7 @@
   @brief Returns the engine type of a SAS library
   @details Usage:
 
-      %put %mf_getEngine(SASHELP);
+      %put %mf_getengine(SASHELP);
 
   returns:
   > V9
@@ -21,9 +21,10 @@
 
   @version 9.2
   @author Allan Bowe
-**/
 
-%macro mf_getEngine(libref
+**/ /** \cond */
+
+%macro mf_getengine(libref
 )/*/STORE SOURCE*/;
   %local dsid engnum rc engine;
 
@@ -42,3 +43,5 @@
  &engine
 
 %mend;
+
+/** \endcond */

base/mp_csv2ds.sas

@@ -0,0 +1,109 @@
+/**
+  @file mp_csv2ds.sas
+  @brief Efficient import of arbitrary CSV using a dataset as template
+  @details Used to import relevant columns from a large CSV using
+  a dataset to provide the types and lengths.  Assumes that a header
+  row is provided, and datarows start on line 2.  Extra columns in
+  both the CSV and base dataset are ignored.
+
+  Usage:
+
+      filename mycsv temp;
+      data _null_;
+        file mycsv;
+        put 'name,age,nickname';
+        put 'John,48,Jonny';
+        put 'Jennifer,23,Jen';
+      run;
+
+      %mp_csv2ds(inref=mycsv,outds=myds,baseds=sashelp.class)
+
+
+  @param inref= fileref to the CSV
+  @param outds= output ds.  Could also be a view (eg `outds=myds/view=myds`)
+  @param baseds= Template dataset on which to create the input statement.
+    Is used to determine types, lengths, and any informats.
+
+  @version 9.2
+  @author Allan Bowe
+**/
+
+%macro mp_csv2ds(inref=0,outds=0,baseds=0);
+%if &inref=0 %then %do;
+  %put %str(ERR)OR: the INREF variable must be provided;
+  %let syscc=4;
+  %abort;
+%end;
+%if &outds=0 %then %do;
+  %put %str(ERR)OR: the OUTDS variable must be provided;
+  %let syscc=4;
+  %return;
+%end;
+%if &baseds=0 %then %do;
+  %put %str(ERR)OR: the BASEDS variable must be provided;
+  %let syscc=4;
+  %return;
+%end;
+
+/* get the variables in the CSV */
+data _data_;
+  infile &inref;
+  input;
+  length name $32;
+  do i=1 to countc(_infile_,',')+1;
+    name=upcase(scan(_infile_,i,','));
+    output;
+  end;
+  stop;
+run;
+%local csv_vars;%let csv_vars=&syslast;
+
+/* get the variables in the dataset */
+proc contents noprint data=&baseds
+  out=_data_ (keep=name type length format: informat);
+run;
+%local base_vars; %let base_vars=&syslast;
+
+proc sql undo_policy=none;
+create table &csv_vars as
+  select a.*
+    ,b.type
+    ,b.length
+    ,b.format
+    ,b.formatd
+    ,b.formatl
+    ,b.informat
+  from &csv_vars a
+  left join &base_vars b
+  on a.name=upcase(b.name)
+  order by i;
+
+/* prepare the input statement */
+%local instat dropvars;
+data _null_;
+  set &syslast end=last;
+  length in dropvars $32767;
+  retain in dropvars;
+  if missing(type) then do;
+    informat='$1.';
+    dropvars=catx(' ',dropvars,name);
+  end;
+  else if missing(informat) then do;
+    if type=1 then informat='best.';
+    else informat=cats('$',length,'.');
+  end;
+  in=catx(' ',in,name,':',informat);
+  if last then do;
+    call symputx('instat',in,'l');
+    call symputx('dropvars',dropvars,'l');
+  end;
+run;
+
+/* import the CSV */
+data &outds;
+  infile &inref dsd firstobs=2;
+  input &instat;
+  drop &dropvars;
+run;
+
+%mend;

base/mp_dirlist.sas

@@ -85,7 +85,8 @@ data &outds (compress=no keep=file_or_folder filepath filename ext msg directory
   dnum = dnum(did);
   do i = 1 to dnum;
     filename = dread(did, i);
-    rc = filename(fref2, cats(directory,'/',filename));
+    filepath=cats(directory,'/',filename);
+    rc = filename(fref2,filepath);
     midd=dopen(fref2);
     dmsg=sysmsg();
     if did > 0 then file_or_folder='folder';
@@ -108,7 +109,6 @@ data &outds (compress=no keep=file_or_folder filepath filename ext msg directory
       ext='';
       file_or_folder='folder';
     end;
-    filepath="&path/"!!filename;
     output;
   end;
   rc = dclose(did);

base/mp_getddl.sas

@@ -30,7 +30,6 @@
    datetime2 format or regular decimal type
   @version 9.3
   @author Allan Bowe
-  @source https://github.com/sasjs/core
 **/
 
 %macro mp_getddl(libref,ds,fref=getddl,flavour=SAS,showlog=NO,schema=
@@ -132,7 +131,7 @@ run;
     %let curds=%scan(&dsnlist,&x);
     data _null_;
       file &fref mod;
-      length nm lab $1024;
+      length nm lab $1024 typ $20;
       set &colinfo (where=(upcase(memname)="&curds")) end=last;
 
       if _n_=1 then do;
@@ -146,10 +145,12 @@ run;
       end;
       else put "   ,"@@;
       if length(format)>1 then fmt=" format="!!cats(format);
-      len=" length="!!cats(length);
+      if length(label)>1 then lab=" label="!!quote(trim(label));
-      lab=" label="!!quote(trim(label));
       if notnull='yes' then notnul=' not null';
-      put name type len fmt notnul lab;
+      if type='char' then typ=cats('char(',length,')');
+      else if length ne 8 then typ='num length='!!left(length);
+      else typ='num';
+      put name typ fmt notnul lab;
     run;
 
     /* Extra step for data constraints */