Parsing flat text with format field="value" in QlikView SQL

[#1]

I'm bumping into QlikView's limited ETL functionality while looking for a way to parse the following data format :

field1=value1 field2="value 2" field3=value

field1=value2 field3=value

field1=value3 field2="another value" field4="yet another value"

-> All records are one line

-> Values with more than one word are quoted

-> Not all lines have all values, so just using subfield() is not going to do it

-> If it's not doable in QlikView script alone (I found no solution to this problem to date) I can live with an external script that can be triggered at load that "normalizes" the data in a format QlikView can parse correctly

Looking forward to your input.

[Gysbert_Wassenaar]

Table:

GENERIC

LOAD ID, SubField(FieldValue,'=',1) as Field,  SubField(FieldValue,'=',-1) as Value;

LOAD RecNo() as ID, subfield(replace(Line, ' field','█field'),'█') as FieldValue INLINE [

  Line

    'field1=value1 field2="value 2" field3=value'

    field1=value2 field3=value

    'field1=value3 field2="another value" field4="yet another value"'

];

And if you think you can live only with one resulting table you can read this blog post: The Generic Load

[#3]

Surprisingly elegant solution Gysbert, I can almost wrap my head around it.

However, I'm not 100% how this would work when :

-> I'm not working with an inline table but I load the data from a file on disk

-> field1, field2, field3 are not the actual field names but can be any descriptive text string that is one word.