node.js - Streaming big data while sorting -
i have huge data , result cannot hold of in memory , out of memory errors; 1 of solutions using streaming in node.js; streaming not possible(as far know) sorting 1 functionalities apply on data; there algorithm maybe divide , conquer algorithm can use combination of streaming , sorting (which 1 of functionalities apply on data?)
you can stream data using kinesis , use kinesis client library, or subscribe lambda function kinesis stream , incrementally maintain sorted materialized views. store sorted materialized views , how divide data depend on application. if cannot store entire sorted materialized views, have rolling views. if data time-series, or has other natural order, divide range of ordered attribute chunks. then, have example, 1-day or 1-hour sorted chunks of data. in other words, choose sorted subdivision allows keep information in memory needed.
Comments
Post a Comment