topic Re: Debug URL Return for PubMed Search EUTILS>NCBI in SAS Programming

Debug URL Return for PubMed Search EUTILS>NCBI

bhallissey — Sat, 31 Aug 2019 03:27:33 GMT

This code used to work and no longer does. I'm not familiar with parsing/accessing url results so cannot even figure out how to debug. Can anyone offer any insights to see what is being returned, to debug, to fix?! Thank you.

filename y url "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=PNAS[ta]+AND+97[vi]&retstart=6&retmax=6&tool=biomed3";

data pubmedinfo;
length countfound count loc locend 8. countchar $30.;
retain count countfound 0 countchar "TTTT";
infile y lrecl=32000 pad;
input;
if countfound = 0 then do;
  loc = find(_infile_,'<Count>');
  locend = find(_infile_,'</Count>');
end;
if loc not in (0 .) then do;
   countchar = substr(_infile_,loc+7,locend - (loc+7));
   count = countchar;
   countfound = 1;
   output;
end;
run;

Re: Debug URL Return for PubMed Search EUTILS>NCBI

VDD — Sat, 31 Aug 2019 14:17:38 GMT

your statement that it used to work and no longer does is not clear. How do you know that it was working if you are unfamiliar with parsing?

Re: Debug URL Return for PubMed Search EUTILS>NCBI

Tom — Sat, 31 Aug 2019 18:32:56 GMT

What does working mean?

If I try to open that URL I get this XML displayed. Notice that it has multiple <COUNT></COUNT> tags.

<eSearchResult>
<Count>2651</Count>
<RetMax>6</RetMax>
<RetStart>6</RetStart>
<IdList>
<Id>11121077</Id>
<Id>11121076</Id>
<Id>11121075</Id>
<Id>11121074</Id>
<Id>11121073</Id>
<Id>11121072</Id>
</IdList>
<TranslationSet>
<Translation>
<From>PNAS[ta]</From>
<To>"Proc Natl Acad Sci U S A"[Journal]</To>
</Translation>
</TranslationSet>
<TranslationStack>
<TermSet>
<Term>"Proc Natl Acad Sci U S A"[Journal]</Term>
<Field>Journal</Field>
<Count>137446</Count>
<Explode>N</Explode>
</TermSet>
<TermSet>
<Term>97[vi]</Term>
<Field>vi</Field>
<Count>93303</Count>
<Explode>N</Explode>
</TermSet>
<OP>AND</OP>
</TranslationStack>
<QueryTranslation>"Proc Natl Acad Sci U S A"[Journal] AND 97[vi]</QueryTranslation>
</eSearchResult>

If I read it with SAS data step I see that it is actually only 10 list of text. The Browser was just breaking it into multiple lines to make it easier for humans to look at.

543   data _null_;
544     infile y ;
545     input ;
546     put _infile_;
547   run;

NOTE: The infile Y is:

      Filename=https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=PNAS[ta]+AND+97[vi
      ]&retstart=6&retmax=6&tool=biomed3,
      Local Host Name=AMRL20L6F1E4992,
      Local Host IP addr=fe80::c856:15a5:8ebd:8fdc%7,
      Service Hostname Name=www.ncbi.nlm.nih.gov,
      Service IP addr=130.14.29.110,Service Name=N/A,
      Service Portno=443,Lrecl=32767,Recfm=Variable

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE eSearchResult PUBLIC "-//NLM//DTD esearch 20060628//EN" "https://eutils.ncbi.nlm.nih.gov/eutils/d
td/20060628/esearch.dtd">
<eSearchResult><Count>2651</Count><RetMax>6</RetMax><RetStart>6</RetStart><IdList>
<Id>11121077</Id>
<Id>11121076</Id>
<Id>11121075</Id>
<Id>11121074</Id>
<Id>11121073</Id>
<Id>11121072</Id>
</IdList><TranslationSet><Translation>     <From>PNAS[ta]</From>     <To>"Proc Natl Acad Sci U S A"[Journal
]</To>    </Translation></TranslationSet><TranslationStack>   <TermSet>    <Term>"Proc Natl Acad Sci U S A"
[Journal]</Term>    <Field>Journal</Field>    <Count>137446</Count>    <Explode>N</Explode>   </TermSet>
<TermSet>    <Term>97[vi]</Term>    <Field>vi</Field>    <Count>93303</Count>    <Explode>N</Explode>   </T
ermSet>   <OP>AND</OP>  </TranslationStack><QueryTranslation>"Proc Natl Acad Sci U S A"[Journal] AND 97[vi]
</QueryTranslation></eSearchResult>
NOTE: 10 records were read from the infile Y.
      The minimum record length was 17.
      The maximum record length was 570.

And if I run your program it "works" and finds the first COUNT value.

596   data pubmedinfo;
597   length countfound count loc locend 8. countchar $30.;
598   retain count countfound 0 countchar "TTTT";
599   infile y lrecl=32000 pad;
600   input;
601        if countfound = 0 then do;
602             loc = find(_infile_,'<Count>');
603             locend = find(_infile_,'</Count>');
604        end;
605        if loc not in (0 .) then do;
606             countchar = substr(_infile_,loc+7,locend - (loc+7));
607             count = countchar;
608             countfound = 1;
609             output;
610        end;
611   run;

NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column).
      607:19
NOTE: The infile Y is:

      Filename=https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=PNAS[ta]+AND+97[vi
      ]&retstart=6&retmax=6&tool=biomed3,
      Local Host Name=AMRL20L6F1E4992,
      Local Host IP addr=fe80::c856:15a5:8ebd:8fdc%7,
      Service Hostname Name=www.ncbi.nlm.nih.gov,
      Service IP addr=130.14.29.110,Service Name=N/A,
      Service Portno=443,Lrecl=32000,Recfm=Variable

NOTE: 10 records were read from the infile Y.
      The minimum record length was 17.
      The maximum record length was 570.
NOTE: The data set WORK.PUBMEDINFO has 1 observations and 5 variables.
NOTE: DATA statement used (Total process time):
      real time           0.19 seconds
      cpu time            0.04 seconds


612
613   data _null_;
614    set pubmedinfo;
615    put (_all_) (=/);
616   run;


countfound=1
count=2651
loc=16
locend=27
countchar=2651
NOTE: There were 1 observations read from the data set WORK.PUBMEDINFO.
NOTE: DATA statement used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds

What do you want to do differently?

Re: Debug URL Return for PubMed Search EUTILS>NCBI

Tom — Sat, 31 Aug 2019 18:32:02 GMT

Try this instead.

data pubmedinfo;
  infile y dlm='<';
  input @'<Count>' @;
  input count @@ ;
run;

data _null_;
 set pubmedinfo;
 put (_all_) (=/);
run;

count=2651

count=137446

count=93303
NOTE: There were 3 observations read from the data set WORK.PUBMEDINFO.

Re: Debug URL Return for PubMed Search EUTILS>NCBI

Tom — Sat, 31 Aug 2019 19:03:02 GMT

Note:

Don't place URL's like that with &'s in them inside of double quotes. The SAS macro processor will try to convert anything that looks like a macro variable reference, like &term into the value of the macro variable. Use single quotes instead and the macro processor will ignore the &'s and %'s in the string.

If you need to build the URL from pieces then do so in a data step. You could then use the FILENAME() function to generate your fileref.

data _null_;
  url='https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=PNAS[ta]+AND+97[vi]&retstart=6&retmax=6&tool=biomed3';
  rc=filename('Y',url,'url');
run;