database - Python Keeping track of changes -

i'm trying identify new, , terminated transaction month month using python 2.7.

i have .csv can 2-3gb large containing 4 months i'm looking solution quick possible instead of putting large dictionaries in memory , iterating on them. in csv data looks this:

date,uniqueid,highlevelclientdetail

i'm wondering how output summary statistics each month of total number (continued uniqueid, terminated, new). edit: continued unique id refers id appears in month 1 , 2, terminated means appears in month 1 not 2, , new refers being in month 2 not month 1

normally, i'd put statistical software , 2 lines i'm limited using python , have never used in way.