@sasjs/utils 2.48.2 → 2.49.1

package/macros/mp_ds2inserts.sas

@@ -0,0 +1,180 @@
+/**
+  @file
+  @brief Export a dataset to SQL insert statements
+  @details Converts dataset values to SQL insert statements for use across
+  multiple database types.
+
+  Usage:
+
+      %mp_ds2inserts(sashelp.class,outref=myref,outds=class)
+      data class;
+        set sashelp.class;
+        stop;
+      proc sql;
+      %inc myref;
+
+  @param [in] ds The dataset to be exported
+  @param [in] maxobs= (max) The max number of inserts to create
+  @param [out] outref= (0) The output fileref.  If it does not exist, it is
+    created. If it does exist, new records are APPENDED.
+  @param [out] schema= (0) The library (or schema) in which the target table is
+    located.  If not provided, is ignored.
+  @param [out] outds= (0) The output table to load.  If not provided, will
+    default to the table in the &ds parameter.
+  @param [in] flavour= (SAS) The SQL flavour to be applied to the output. Valid
+    options:
+    @li SAS (default) - suitable for regular proc sql
+    @li PGSQL - Used for Postgres databases
+  @param [in] applydttm= (YES) If YES, any columns using datetime formats will
+    be converted to native DB datetime literals
+
+  <h4> SAS Macros </h4>
+  @li mf_existfileref.sas
+  @li mf_getvarcount.sas
+  @li mf_getvarformat.sas
+  @li mf_getvarlist.sas
+  @li mf_getvartype.sas
+
+  @version 9.2
+  @author Allan Bowe (credit mjsq)
+**/
+
+%macro mp_ds2inserts(ds, outref=0,schema=0,outds=0,flavour=SAS,maxobs=max
+  ,applydttm=YES
+)/*/STORE SOURCE*/;
+
+%if not %sysfunc(exist(&ds)) %then %do;
+  %put %str(WAR)NING:  &ds does not exist;
+  %return;
+%end;
+
+%if not %sysfunc(exist(&ds)) %then %do;
+  %put %str(WAR)NING:  &ds does not exist;
+  %return;
+%end;
+
+%if %index(&ds,.)=0 %then %let ds=WORK.&ds;
+
+%let flavour=%upcase(&flavour);
+%if &flavour ne SAS and &flavour ne PGSQL %then %do;
+  %put %str(WAR)NING:  &flavour is not supported;
+  %return;
+%end;
+
+%if &outref=0 %then %do;
+  %put %str(WAR)NING:  Please provide a fileref;
+  %return;
+%end;
+%if %mf_existfileref(&outref)=0 %then %do;
+  filename &outref temp lrecl=66000;
+%end;
+
+%if &schema=0 %then %let schema=;
+%else %let schema=&schema..;
+
+%if &outds=0 %then %let outds=%scan(&ds,2,.);
+
+%local nobs;
+proc sql noprint;
+select count(*) into: nobs TRIMMED from &ds;
+%if &nobs=0 %then %do;
+  data _null_;
+    file &outref mod;
+    put "/* No rows found in &ds */";
+  run;
+%end;
+
+%local vars;
+%let vars=%mf_getvarcount(&ds);
+%if &vars=0 %then %do;
+  data _null_;
+    file &outref mod;
+    put "/* No columns found in &schema.&ds */";
+  run;
+  %return;
+%end;
+%else %if &vars>1600 and &flavour=PGSQL %then %do;
+  data _null_;
+    file &fref mod;
+    put "/* &schema.&ds contains &vars vars */";
+    put "/* Postgres cannot handle tables with over 1600 vars */";
+    put "/* No inserts will be generated for this table */";
+  run;
+  %return;
+%end;
+
+%local varlist varlistcomma;
+%let varlist=%mf_getvarlist(&ds);
+%let varlistcomma=%mf_getvarlist(&ds,dlm=%str(,),quote=double);
+
+/* next, export data */
+data _null_;
+  file &outref mod ;
+  if _n_=1 then put "/* &schema.&outds (&nobs rows, &vars columns) */";
+  set &ds;
+  %if &maxobs ne max %then %do;
+    if _n_>&maxobs then stop;
+  %end;
+  length _____str $32767;
+  call missing(_____str);
+  format _numeric_ best.;
+  format _character_ ;
+  %local i comma var vtype vfmt;
+  %do i=1 %to %sysfunc(countw(&varlist));
+    %let var=%scan(&varlist,&i);
+    %let vtype=%mf_getvartype(&ds,&var);
+    %let vfmt=%upcase(%mf_getvarformat(&ds,&var,force=1));
+    %if &i=1 %then %do;
+      %if &flavour=SAS %then %do;
+        put "insert into &schema.&outds set ";
+        put "  &var="@;
+      %end;
+      %else %if &flavour=PGSQL %then %do;
+        _____str=cats(
+          "INSERT INTO &schema.&outds ("
+          ,symget('varlistcomma')
+          ,") VALUES ("
+        );
+        put _____str;
+        put "  "@;
+      %end;
+    %end;
+    %else %do;
+      %if &flavour=SAS %then %do;
+        put "  ,&var="@;
+      %end;
+      %else %if &flavour=PGSQL %then %do;
+        put "  ,"@;
+      %end;
+    %end;
+    %if &vtype=N %then %do;
+      %if &flavour=SAS %then %do;
+        put &var;
+      %end;
+      %else %if &flavour=PGSQL %then %do;
+        if missing(&var) then put 'NULL';
+        %if &applydttm=YES and "%substr(&vfmt.xxxxxxxx,1,8)"="DATETIME"
+        %then %do;
+          else put "TIMESTAMP '" &var E8601DT25.6 "'";
+        %end;
+        %else %do;
+          else put &var;
+        %end;
+      %end;
+    %end;
+    %else %do;
+      _____str="'"!!trim(tranwrd(&var,"'","''"))!!"'";
+      put _____str;
+    %end;
+  %end;
+  %if &flavour=SAS %then %do;
+    put ';';
+  %end;
+  %else %if &flavour=PGSQL %then %do;
+    put ');';
+  %end;
+
+  if _n_=&nobs then put /;
+run;
+
+%mend mp_ds2inserts;

package/macros/mp_ds2md.sas

@@ -0,0 +1,103 @@
+/**
+  @file
+  @brief Create a Markdown Table from a dataset
+  @details A markdown table is a simple table representation for use in
+  documents written in markdown format.
+
+  An online generator is available here:
+  https://www.tablesgenerator.com/markdown_tables
+
+  This structure is also used by the Macro Core library for documenting input/
+  output datasets, as well as the sasjs/cli tool for documenting inputs/outputs
+  for web services.
+
+  We take the standard definition one step further by embedding the informat
+  in the table header row, like so:
+
+      |var1:$32|var2:best.|var3:date9.|
+      |---|---|---|
+      |some text|42|01JAN1960|
+      |blah|1|31DEC1999|
+
+  Which resolves to:
+
+  |var1:$32|var2:best.|var3:date9.|
+  |---|---|---|
+  |some text|42|01JAN1960|
+  |blah|1|31DEC1999|
+
+
+  Usage:
+
+      %mp_ds2md(sashelp.class)
+
+  @param [in] libds the library / dataset to create or read from.
+  @param [out] outref= (mdtable) Fileref to contain the markdown
+  @param [out] showlog= (YES) Set to NO to avoid printing markdown to the log
+
+  <h4> SAS Macros </h4>
+  @li mf_getvarlist.sas
+  @li mf_getvarformat.sas
+
+  @version 9.3
+  @author Allan Bowe
+**/
+
+%macro mp_ds2md(
+  libds,
+  outref=mdtable,
+  showlog=YES
+)/*/STORE SOURCE*/;
+
+/* check fileref is assigned */
+%if %sysfunc(fileref(&outref)) > 0 %then %do;
+  filename &outref temp;
+%end;
+
+%local vars;
+%let vars=%upcase(%mf_getvarlist(&libds));
+
+%if %trim(X&vars)=X %then %do;
+  %put &sysmacroname: Table &libds has no columns!!;
+  %return;
+%end;
+
+/* create the header row */
+data _null_;
+  file &outref;
+  length line $32767;
+  call missing(line);
+  put '|'
+%local i var fmt;
+%do i=1 %to %sysfunc(countw(&vars));
+  %let var=%scan(&vars,&i);
+  %let fmt=%lowcase(%mf_getvarformat(&libds,&var,force=1));
+  "&var:&fmt|"
+%end;
+  ;
+  put '|'
+%do i=1 %to %sysfunc(countw(&vars));
+  "---|"
+%end;
+  ;
+run;
+
+/* write out the data */
+data _null_;
+  file &outref mod dlm='|' lrecl=32767;
+  set &libds ;
+  length line $32767;
+  line='|`'!!cats(%mf_getvarlist(&libds,dlm=%str(%)!!' `|`'!!cats%()))!!' `|';
+  put line;
+run;
+
+%if %upcase(&showlog)=YES %then %do;
+  options ps=max lrecl=max;
+  data _null_;
+    infile &outref;
+    input;
+    putlog _infile_;
+  run;
+%end;
+
+%mend mp_ds2md;

package/macros/mp_ds2squeeze.sas

@@ -0,0 +1,120 @@
+/**
+  @file
+  @brief Create a smaller version of a dataset, without data loss
+  @details This macro will scan the input dataset and create a new one, that
+  has the minimum variable lengths needed to store the data without data loss.
+
+  Inspiration was taken from [How to Reduce the Disk Space Required by a
+  SAS® Data Set](https://www.lexjansen.com/nesug/nesug06/io/io18.pdf) by
+  Selvaratnam Sridharma.  The end of the referenced paper presents a macro named
+  "squeeze", hence the nomenclature.
+
+  Usage:
+
+      data big;
+        length my big $32000;
+        do i=1 to 1e4;
+          my=repeat('oh my',100);
+          big='dawg';
+          special=._;
+          output;
+        end;
+      run;
+
+      %mp_ds2squeeze(work.big,outds=work.smaller)
+
+  The following will also be printed to the log (exact values may differ
+  depending on your OS and COMPRESS settings):
+
+  > MP_DS2SQUEEZE: work.big was  625MB
+
+  > MP_DS2SQUEEZE: work.smaller is    5MB
+
+  @param [in] libds The library.dataset to be squeezed
+  @param [out] outds= (work.mp_ds2squeeze) The squeezed dataset to create
+  @param [in] mdebug= (0) Set to 1 to enable DEBUG messages
+
+  <h4> SAS Macros </h4>
+  @li mf_getfilesize.sas
+  @li mf_getuniquefileref.sas
+  @li mf_getuniquename.sas
+  @li mp_getmaxvarlengths.sas
+
+  <h4> Related Programs </h4>
+  @li mp_ds2squeeze.test.sas
+
+  @version 9.3
+  @author Allan Bowe
+**/
+
+%macro mp_ds2squeeze(
+  libds,
+  outds=work.mp_ds2squeeze,
+  mdebug=0
+)/*/STORE SOURCE*/;
+%local dbg source;
+%if &mdebug=1 %then %do;
+  %put &sysmacroname entry vars:;
+  %put _local_;
+%end;
+%else %do;
+  %let dbg=*;
+  %let source=/source2;
+%end;
+
+%local optval ds fref startsize;
+%let ds=%mf_getuniquename();
+%let fref=%mf_getuniquefileref();
+%let startsize=%mf_getfilesize(libds=&libds,format=yes);
+
+%mp_getmaxvarlengths(&libds,outds=&ds)
+
+data _null_;
+  set &ds end=last;
+  file &fref;
+  /* grab the types */
+  retain dsid;
+  if _n_=1 then dsid=open("&libds",'is');
+  if dsid le 0 then do;
+    msg=sysmsg();
+    put msg=;
+    stop;
+  end;
+  type=vartype(dsid,varnum(dsid, name));
+  if last then rc=close(dsid);
+  /* write out the length statement */
+  if _n_=1 then put 'length ';
+  length len $6;
+  if type='C' then do;
+    if maxlen=0 then len='$1';
+    else len=cats('$',maxlen);
+  end;
+  else do;
+    if maxlen=0 then len='3';
+    else len=cats(maxlen);
+  end;
+  put '  ' name ' ' len;
+  if last then put ';';
+run;
+
+/* configure varlenchk - as we are explicitly shortening the variables */
+%let optval=%sysfunc(getoption(varlenchk));
+options varlenchk=NOWARN;
+
+data &outds;
+  %inc &fref &source;
+  set &libds;
+run;
+
+options varlenchk=&optval;
+
+%if &mdebug=0 %then %do;
+  proc sql;
+  drop table &ds;
+  filename &fref clear;
+%end;
+
+%put &sysmacroname: &libds was &startsize;
+%put &sysmacroname: &outds is %mf_getfilesize(libds=&outds,format=yes);
+
+%mend mp_ds2squeeze;

package/macros/mp_dsmeta.sas

@@ -0,0 +1,110 @@
+/**
+  @file
+  @brief Export dataset metadata to a single output table
+  @details Exports the dataset attributes and enginehost information, then
+  converts the datasets into a single output table in the following format:
+
+|ODS_TABLE:$10.|NAME:$100.|VALUE:$1000.|
+|---|---|---|
+|`ATTRIBUTES `|`Data Set Name `|`SASHELP.CLASS `|
+|`ATTRIBUTES `|`Observations `|`19 `|
+|`ATTRIBUTES `|`Member Type `|`DATA `|
+|`ATTRIBUTES `|`Variables `|`5 `|
+|`ATTRIBUTES `|`Engine `|`V9 `|
+|`ATTRIBUTES `|`Indexes `|`0 `|
+|`ATTRIBUTES `|`Created `|`06/08/2020 00:59:14 `|
+|`ATTRIBUTES `|`Observation Length `|`40 `|
+|`ATTRIBUTES `|`Last Modified `|`06/08/2020 00:59:14 `|
+|`ATTRIBUTES `|`Deleted Observations `|`0 `|
+|`ATTRIBUTES `|`Protection `|`. `|
+|`ATTRIBUTES `|`Compressed `|`NO `|
+|`ATTRIBUTES `|`Data Set Type `|`. `|
+|`ATTRIBUTES `|`Sorted `|`NO `|
+|`ATTRIBUTES `|`Label `|`Student Data `|
+|`ATTRIBUTES `|`Data Representation `|`SOLARIS_X86_64, LINUX_X86_64, ALPHA_TRU64, LINUX_IA64 `|
+|`ATTRIBUTES `|`Encoding `|`us-ascii  ASCII (ANSI) `|
+|`ENGINEHOST `|`Data Set Page Size `|`65536 `|
+|`ENGINEHOST `|`Number of Data Set Pages `|`1 `|
+|`ENGINEHOST `|`First Data Page `|`1 `|
+|`ENGINEHOST `|`Max Obs per Page `|`1632 `|
+|`ENGINEHOST `|`Obs in First Data Page `|`19 `|
+|`ENGINEHOST `|`Number of Data Set Repairs `|`0 `|
+|`ENGINEHOST `|`Filename `|`/opt/sas/sas9/SASHome/SASFoundation/9.4/sashelp/class.sas7bdat `|
+|`ENGINEHOST `|`Release Created `|`9.0401M7 `|
+|`ENGINEHOST `|`Host Created `|`Linux `|
+|`ENGINEHOST `|`Inode Number `|`28314616 `|
+|`ENGINEHOST `|`Access Permission `|`rw-r--r-- `|
+|`ENGINEHOST `|`Owner Name `|`sas `|
+|`ENGINEHOST `|`File Size `|`128KB `|
+|`ENGINEHOST `|`File Size (bytes) `|`131072 `|
+
+  Example usage:
+
+      %mp_dsmeta(sashelp.class,outds=work.mymeta)
+      proc print data=work.mymeta;
+      run;
+
+  For more details on creating datasets from PROC CONTENTS check out this
+  excellent [paper](
+  https://support.sas.com/resources/papers/proceedings14/1549-2014.pdf) by
+  [Louise Hadden](https://www.linkedin.com/in/louisehadden/).
+
+  @param libds The library.dataset to export the metadata for
+  @param outds= (work.dsmeta) The output table to contain the metadata
+
+  <h4> Related Files </h4>
+  @li mp_dsmeta.test.sas
+  @li mp_getcols.sas
+  @li mp_getdbml.sas
+  @li mp_getddl.sas
+  @li mp_getformats.sas
+  @li mp_getpk.sas
+  @li mp_guesspk.sas
+
+**/
+
+%macro mp_dsmeta(libds,outds=work.dsmeta);
+
+%local ds1 ds2;
+data;run; %let ds1=&syslast;
+data;run; %let ds2=&syslast;
+
+/* setup the ODS capture */
+ods output attributes=&ds1 enginehost=&ds2;
+
+/* export the metadata */
+proc contents data=&libds;
+run;
+
+/* load it into a single table */
+data &outds (keep=ods_table name value);
+  length ods_table $10 name label2 label1 label $100
+    value cvalue cvalue1 cvalue2 $1000
+    nvalue nvalue1 nvalue2 8;
+  if _n_=1 then call missing (of _all_);
+  * putlog (_all_)(=);
+  set &ds1 (in=atrs) &ds2 (in=eng);
+  if atrs then do;
+    ods_table='ATTRIBUTES';
+    name=coalescec(label1,label);
+    value=coalescec(cvalue1,cvalue,put(coalesce(nvalue1,nvalue),best.));
+    output;
+    if label2 ne '' then do;
+    name=label2;
+    value=coalescec(cvalue2,put(nvalue2,best.));
+    output;
+    end;
+  end;
+  else if eng then do;
+    ods_table='ENGINEHOST';
+    name=coalescec(label1,label);
+    value=coalescec(cvalue1,cvalue,put(coalesce(nvalue1,nvalue),best.));
+    output;
+  end;
+run;
+
+proc sql;
+drop table &ds1, &ds2;
+
+%mend mp_dsmeta;
+