I'm running a Spark Job in Scala and I'm struck with parsing the input file.
The Input file(TAB separated) is something like,
date=20160701 name=mike age=26
date=20160402 name=john age=33
I want to parse it and extract only values and not the keys, such as,
20160701 mike 26
20160402 john 33
How can this be achieved in SCALA?
I'm using,
SCALA VERSION: 2.11
You can use CSVParser() and you know the location for key, it will be easy and clean
Test data
val data = "date=20160701\tname=mike\tage=26\ndate=20160402 name=john\tage=33\n"
One statement to do what you asked
val rdd = sc.parallelize(data.split('\n'))
.map(_.split('\t') // split into key=value
.map(_.split('=')(1))) // split those at "=" and select only the value
Display what we got
rdd.collect().foreach(r=>println(r.mkString(",")))
// 20160701,mike,26
// 20160402,john,33
But don't do this for real code. It's very fragile in the face of data format errors, etc. Use CSVParser or something instead as Narendra Parmar suggests.
val rdd = sc.textFile()
rdd.map(x => x.split("\t")).map(x => x.split("=")(1)).map(x => x.mkstring("\t")).saveAsTextFile("")
Related
I want to remove header from a file. But, since the file will be split into partitions, I can't just drop the first item. So I was using a filter function to figure it out and here below is the code I am using :
val noHeaderRDD = baseRDD.filter(line=>!line.contains("REPORTDATETIME"));
and the error I am getting says "error not found value line "what could be the issue here with this code?
I don't think anybody answered the obvious, whereby line.contains also possible:
val noHeaderRDD = baseRDD.filter(line => !(line contains("REPORTDATETIME")))
You were nearly there, just a syntax issue, but that is significant of course!
Using textFile as below:
val rdd = sc.textFile(<<path>>)
rdd.filter(x => !x.startsWith(<<"Header Text">>))
Or
In Spark 2.0:
spark.read.option("header","true").csv("filePath")
This is my first time using Scala and ApacheSpark for a project. I'm trying to print the contents of an matrix when I run my code in the terminal, but nothing I try is working so far.
Instead I only get this printed:
org.apache.spark.mllib.linalg.distributed.MatrixEntry;#71870da7
org.apache.spark.mllib.linalg.distributed.CoordinateMatrix#1dcca8d3
I just using println() but when I use collect(), that doesn't give a good result either.
The default toString prints the name of a class followed by an address in memory.
org.apache.spark.mllib.linalg.distributed.MatrixEntry;#71870da7
You're going to want to find a way to iterate through your matrix and print each element.
Building on #zero323 's comment ( aside would you like to put an answer out there?): given an RDD[SomeType] you can call
rdd.collect()
or
rdd.take(k)
Then you can print out the results using normal toString() methods that depend on the type of the rdd contents. So if SomeType were a List[Double] then the
println(s"${rdd.collect().mkString(",")}")
would give you a single-line comma separated output of the results.
As #zero323 another consideration is: "do you really want to print out the contents of your rdd?" More likely you might only want a summary - such as
println(s"Number of entries in RDD is ${rdd.count()}")
Iterate over the rdd like this,
rdd.foreach(println)
scala>val rdd1 = sc.parallelize(List(1,2,3,4)).map(_*2)
To print the data within RDD
scala> rdd1.collect().foreach(println)
Output:
2
4
6
8
I have developed a hadoop based solution that process a binary file. This uses classic hadoop MR technique. The binary file is about 10GB and divided into 73 HDFS blocks, and the business logic written as map process operates on each of these 73 blocks. We have developed a customInputFormat and CustomRecordReader in Hadoop that returns key (intWritable) and value (BytesWritable) to the map function. The value is nothing but the contents of a HDFS block(bianry data). The business logic knows how to read this data.
Now, I would like to port this code in spark. I am a starter in spark and could run simple examples (wordcount, pi example) in spark. However, could not straightforward example to process binaryFiles in spark. I see there are two solutions for this use case. In the first, avoid using custom input format and record reader. Find a method (approach) in spark the creates a RDD for those HDFS blocks, use a map like method that feeds HDFS block content to the business logic. If this is not possible, I would like to re-use the custom input format and custom reader using some methods such as HadoopAPI, HadoopRDD etc. My problem:- I do not know whether the first approach is possible or not. If possible, can anyone please provide some pointers that contains examples? I was trying second approach but highly unsuccessful. Here is the code snippet I used
package org {
object Driver {
def myFunc(key : IntWritable, content : BytesWritable):Int = {
println(key.get())
println(content.getSize())
return 1
}
def main(args: Array[String]) {
// create a spark context
val conf = new SparkConf().setAppName("Dummy").setMaster("spark://<host>:7077")
val sc = new SparkContext(conf)
println(sc)
val rd = sc.newAPIHadoopFile("hdfs:///user/hadoop/myBin.dat", classOf[RandomAccessInputFormat], classOf[IntWritable], classOf[BytesWritable])
val count = rd.map (x => myFunc(x._1, x._2)).reduce(_+_)
println("The count is *****************************"+count)
}
}
}
Please note that the print statement in the main method prints 73 which is the number of blocks whereas the print statements inside the map function prints 0.
Can someone tell where I am doing wrong here? I think I am not using API the right way but failed to find some documentation/usage examples.
A couple of problems at a glance. You define myFunc but call func. Your myFunc has no return type, so you can't call collect(). If your myFunc truly doesn't have a return value, you can do foreach instead of map.
collect() pulls the data in an RDD to the driver to allow you to do stuff with it locally (on the driver).
I have made some progress in this issue. I am now using the below function which does the job
var hRDD = new NewHadoopRDD(sc, classOf[RandomAccessInputFormat],
classOf[IntWritable],
classOf[BytesWritable],
job.getConfiguration()
)
val count = hRDD.mapPartitionsWithInputSplit{ (split, iter) => myfuncPart(split, iter)}.collect()
However, landed up with another error the details of which i have posted here
Issue in accessing HDFS file inside spark map function
15/10/30 11:11:39 WARN scheduler.TaskSetManager: Lost task 0.0 in stage 0.0 (TID 0, 40.221.94.235): java.io.IOException: No FileSystem for scheme: spark
at org.apache.hadoop.fs.FileSystem.getFileSystemClass(FileSystem.java:2584)
at org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:2591)
at org.apache.hadoop.fs.FileSystem.access$200(FileSystem.java:91)
at org.apache.hadoop.fs.FileSystem$Cache.getInternal(FileSystem.java:2630)
I am using scopt to parse command line arguments in scala. I want it to be able to parse options with more than one value. For instance, the range option, if specified, should take exactly two values.
--range 25 45
Coming, from python background, I am basically looking for a way to do the following with scopt instead of python's argparse:
parser.add_argument("--range", default=None, nargs=2, type=float,
metavar=('start', 'end'),
help=(" Foo bar start and stop "))
I dont think minOccurs and maxOccurs solves my problem exactly, nor the key:value example in its help.
Looking at the source code, this is not possible. The Read type class used has a member tuplesToRead, but it doesn't seem to be working when you force it to 2 instead of 1. You will have to make a feature request, I guess, or work around this by using --min 25 --max 45, or --range '25 45' with a custom Read instance that splits this string into two parts. As #roterl noted, this is not a standard way of parsing.
It should be ok if only your values are delimited with something else than a space...
--range 25-45
... although you need to split them manually. Parse it with something like:
opt[String]('r', "range").action { (x, c) =>
val rx = "([0-9]+)\\-([0-9]+)".r
val rx(from, to) = x
c.copy(from = from.toInt, to = to.toInt)
}
// ...
println(s" Got range ${parsedArgs.from}..${parsedArgs.to}")
I'm reading from a text file, parsing each line to JSON and am attempting to print one of the attributes:
val msgData = ssc.textFileStream(dataDir)
val msgs = msgData.map(MessageParser.parse)
msgs.foreach(msg => println(msg.my_attribute))
However, I get the following error on compilation:
value my_attribute is not a member of org.apache.spark.rdd.RDD[com.imgzine.analytics.messages.Message]
What am I missing?
Thanks
Spark Streaming discretizes a stream of data by creating micro-batch containers. Those are called 'DStreams' and contain a collection of RDD's.
Translated to your case, you need to operate on the content of the RDD, not the DStream:
msgs.foreach(rdd => rdd.foreach(elem => println(elem.my_attribute))
DStreams offer a help method to print the first elements (10 I think) of each RDD:
dstream.print()
Of course, that will just invoke .toString on the objects contained in the RDD and print the result. Maybe not what you want with my_attribute as stated in the question.