4 Replies Latest reply: Sep 21, 2015 5:38 AM by Marco Wedel RSS

    Parsing fields from a string

    Yurii Ratushnyi

      Hi Community,


      I need to reconstruct a table from a string. Input is as follows.


      1Field1 : [Old]ValueOld[New]ValueNew Field2 : [Old]Dummy Text Old [New]New Dummy Text New Field3 : [Old]Dummy Text Old [New]New Dummy Text New


      Wanted ouput:


      1Field2Dummy Text OldDummy Text New
      1Field3Dummy Text OldDummy Text New


      There could be any amount of Fields within a string, however their names are always without spaces

      Stuck a little bit with transformations. Thanks for help!

        • Re: Parsing fields from a string
          Marcus Sommer

          Is this Comment-field the definitely format for this string - then in your example is no unique delimiter for the rows? If you have one it would be quite easy to split the string per subfield(string, delimiter) which split the string on the delimiter and repeated the other parts from this record. Quite similar is to use a while-loop whereby iterno() <= substringcount(delimiter). The other string-parts for the old- and new value could be split per subfield(string, delimiter, number) or textbetween() or maybe some nested mid/left/right- and index-functions.


          I think if you haven't such unique delimiter you will need to do it quite similar only that you will need additionally steps within a if-loop where the current record-value with a previous record will be per peek() or previous() compared and the record-values will be adjusted. The worser the datastructure are the more efforts do you need with this comparing, checking and adjusting.


          Maybe the following links are helpful in understanding what I mean:


          Re: Vertical Data Text in Pivot Table

          Re: Line adding ETL


          - Marcus

          • Re: Parsing fields from a string
            Marco Wedel



            one solution could be:




            LOAD ID,
                SubField(SubComment,' : [',1) as FieldName,
                TextBetween(SubComment,'[',']',IterNo())&'Value' as ColNam,
                TextBetween(SubComment&'[',']','[',IterNo()) as ColVal
            While IterNo() <= SubStringCount(SubComment,'[');
            LOAD ID,
                Trim(Mid(Comment, Alt(Previous(Pos)+1,1),RangeSum(Pos,-Previous(Pos)))) as SubComment;
            LOAD *,
                If(IterNo()=SubStringCount(Comment,' : ['),Len(Comment),Index(Left(Comment,Index(Comment,' : [',IterNo()+1)-1),' ',-1)) as Pos
            INLINE "
                ID, Comment
                1, Field1 : [Old]ValueOld[New]ValueNew Field2 : [Old]Dummy Text Old [New]New Dummy Text New Field3 : [Old]Dummy Text Old [New]New Dummy Text New
            While IterNo() <= SubStringCount(Comment,' : [');


            hope this helps