taking a sparse slice of an array

Thank you for the useful responses to 'Big Wide Datasets'

I decided to have a go with the array option and initially it looked very promising

time psql test -c "SELECT sample_id, snp_data[40000] FROM gwas_test" > /dev/null

returned in 5 minutes from a 100000 row table.

The trouble is that it does not scale

time psql test -c "SELECT sample_id, snp_data[40000], snp_data[50000] FROM gwas_test" > /dev/null

takes 10 minutes :-(

Is there anyway to speed this up I can see snp_data[40000:50000] returns values from 40 to 50 thousand, 
is there any way to so something like

snp_data[1,55,1000,40000,50000]

which would just return a 5 element array?

--
Michael