I have the following response from JMeter:
{\"tot_cnt\":{\"tc\":\"41651\",\"rpp\":\"100\",\"pdid\":\"1594\"},\"mg\":[{\"id\":\"683\",\"rsd\":\"6-May-13 \\/ 6
Interim solution:
In your "Regular Expression Extractor", keep the setting "Match No." as "0"
Even tough the ID's will not be unique but it will randomly select different values rather than selecting same value again & again.
Complete solution:
write all the valid "ID" to CSV file in "1st Thread group" using Bean Shell Sampler
& then
In the "2nd Thread Group" run the test using created CSV with unique values for each request.
While analyzing the results you can ignore 1st thread group.
Hope this will help.