- This topic has 3 replies, 1 voice, and was last updated 5 years, 7 months ago by
.
-
Posts
-
-
-
Mapper runs map functions. It takes data in the form of Key, Value pairs. And the output is 0 or more <K, V> pairs. Maps are the individual tasks which transform input records. into a intermediate records.Mapper/Map task counts the word in each document. Mapper output is stored locally.
Mostly there is one map task for each InputSplit (bytes -oriented view of Input) generated by InputFormat.
InputFormat does following jobs:
It validates the input-specification of the job(Map/Reduce job).
It split input files into logical InputSplits (like one input files is divided into lines so every line will be your input split).
Follow the link to learn more about: Mapper in Hadoop -
MapReduce is the computation layer in hadoop.
The Mapper task is to process input data which is present in HDFS.
It receives data in splits.For every split, there is mapper assigned which processes the split data and produces the output which is stored on disk called as intermediate output.
The data accepted by Mapper is in terms of Key, Value pairs from Record Reader.
The processing of Mapper may differ according to File Input Format.The default is a text format for which mapper processes line by line.
For more detail follow: Mapper in Hadoop
-
Mappers are the individual tasks which transform input records into a intermediate output. The transformed intermediate outputs can be completely different from input pair. Mapper understands only data in the key, value pairs . So, input data should be first converted into key, value pairs before passing to the mapper.
The number of map tasks in a map-reduce program is determined by the total number of blocks of the input file
Mapper= {(total data size)/ (input split size)}
-
- You must be logged in to reply to this topic.
