Sorting and summing large amounts of data

I have a data download from a .sql database to Excel. I have written code to refresh this, and it populates an Excel worksheet. I need to sum this data so that drivers (eg alpha-numeric nominal codes) appear once, and another column is summed if a third column meets certain criteria.

Eg - I may have 25 transactions appearing for nominal code 1234abc, 36 transactions for nominal code 4567def.

I want 1234 in one column, abc in another column, 4567 in the first column a row below 1234 and def in the second column a row below abc. I then want to sum whichever of the 25 transactions (column f - column g in my base data sheet) for 1234abc have occurred in an accounting period <=x (column h in my base data sheet), and ditto for each nominal code occurring in my base data sheet. I then want to return these summed amounts in column 'c' of my reporting sheet, against the corresponding nominal code. I have been told that the best way to do this is to use an ADO connection and create a data set in memory. As I am working within one sheet, is this really necessary?