SQL problem.

For our webapps we use sql statements to retrieve data from our iSeries.  I
used to use the * to retrieve all fields for a record.  Thinking this over,
I realized that I regularly need only half of the fields in the database.
Because our records has mostly between 20 and 50 fields, this could be a
huge datatraffic overhead.
Therefore I tested retrieval of data using SQL as follows:
SQL1= 'select * from dbfile where field1=string'
SQL2= 'select field1, field10, field 17 from dbfile where field1=string'

surprisingly SQL2 is running slower than SQL1!  I think this has something
to do with the creation of a new temporary table with only 3 fields by the
SQL-engine, no?