I have the multiple text files.
The total size of them exceeds the largest disk size available to me (~1.5TB)
A spark program reads a single input text file from HDFS. So I need to combine those files into one. (I cannot re-write the program code. I am given only the *.jar file for execution)
Does HDFS have such a capability? How can I achieve this?
What I understood from your question is you want to Concatenate multiple files into one. Here is a solution which might not be the most efficient way of doing it but it works. suppose you have two files: file1 and file2 and you want to get a combined file as ConcatenatedFile
.Here is the script for that.
hadoop fs -cat /hadoop/path/to/file/file1.txt /hadoop/path/to/file/file2.txt | hadoop fs -put - /hadoop/path/to/file/Concatenate_file_Folder/ConcatenateFile.txt
Hope this helps.
HDFS by itself does not provide such capabilities. All out-of-the-box features (like hdfs dfs -text * with pipes or FileUtil's copy methods) use your client server to transfer all data.
In my experience we always used our own written MapReduce jobs to merge many small files in HDFS in distributed way.
So you have two solutions:
Write your own simple MapReduce/Spark job to combine text files with
your format.
Find already implemented solution for such kind of
purposes.
About solution #2: there is the simple project FileCrush for combining text or sequence files in HDFS. It might be suitable for you, check it.
Example of usage:
hadoop jar filecrush-2.0-SNAPSHOT.jar crush.Crush -Ddfs.block.size=134217728 \
--input-format=text \
--output-format=text \
--compress=none \
/input/dir /output/dir 20161228161647
I had a problem to run it without these options (especially -Ddfs.block.size and output file date prefix 20161228161647) so make sure you run it properly.
You can do a pig job:
A = LOAD '/path/to/inputFiles' as (SCHEMA);
STORE A into '/path/to/outputFile';
Doing a hdfs cat and then putting it back to hdfs means, all this data is processed in the client node and will degradate your network
Related
In a cluster of hdfs, i receive multiple files on a daily basis which can be of 3 types :
1) product_info_timestamp
2) user_info_timestamp
3) user_activity_timestamp
The number of files received can be of any number but they will belong to one of these 3 categories only.
I want to merge all the files(after checking whether they are less than 100mb) belonging to one category into a single file.
for eg: 3 files named product_info_* should be merged into one file named product_info.
How do i achieve this?
You can use getmerge toachieve this, but the result will be stored in your local node (edge node), so you need to be sure you have enough space there.
hadoop fs -getmerge /hdfs_path/product_info_* /local_path/product_inf
You can move them back to hdfs with put
hadoop fs -put /local_path/product_inf /hdfs_path
You can use hadoop archive (.har file) or sequence file. It is very simple to use - just google "hadoop archive" or "sequence file".
Another set of commands along the similar lines as suggested by #SCouto
hdfs dfs -cat /hdfs_path/product_info_* > /local_path/product_info_combined.txt
hdfs dfs -put /local_path/product_info_combined.txt /hdfs_path/
How to compress hdfs data to bzip2 using pig such that on decompression it should give the same dir structure which it had initially.I am new to pig.
I tried to compress with bzip2 but it generated many files due to many mappers being spawned and hence reverting back to plain text file(initial form) in the same dir structure becomes difficult.
Just like how in unix if we compress bzip2 using tarball and then after decompression of bzip2.tar gives me exactly same data and folder structure which it had initially.
eg Compression:- tar -cjf compress_folder.tar.bz2 compress_folder/
Decompression:- tar -jtvf compress_folder.tar.bz2
will give exactly same dir st.
Approach 1:
you can try running one reducer to store only 1 file on hdfs. but compromise will be performance here.
set default_parallel 1;
to compress data, set these parameters in pig script , if not tried this way:-
set output.compression.enabled true;
SET mapred.output.compression.codec 'org.apache.hadoop.io.compress.BZip2Codec';
just use JsonStorage while storing file
STORE file INTO '/user/hduser/data/usercount' USING JsonStorage();
Eventually you also want to read data, use TextLoader
data = LOAD '/user/hduser/data/usercount/' USING TextLoader;
Approach 2:
filecrush: file merge utility available at #Mr. github
Do we need to verify checksum after we move files to Hadoop (HDFS) from a Linux server through a Webhdfs ?
I would like to make sure the files on the HDFS have no corruption after they are copied. But is checking checksum necessary?
I read client does checksum before data is written to HDFS
Can somebody help me to understand how can I make sure that source file on Linux system is same as ingested file on Hdfs using webhdfs.
If your goal is to compare two files residing on HDFS, I would not use "hdfs dfs -checksum URI" as in my case it generates different checksums for files with identical content.
In the below example I am comparing two files with the same content in different locations:
Old-school md5sum method returns the same checksum:
$ hdfs dfs -cat /project1/file.txt | md5sum
b9fdea463b1ce46fabc2958fc5f7644a -
$ hdfs dfs -cat /project2/file.txt | md5sum
b9fdea463b1ce46fabc2958fc5f7644a -
However, checksum generated on the HDFS is different for files with the same content:
$ hdfs dfs -checksum /project1/file.txt
0000020000000000000000003e50be59553b2ddaf401c575f8df6914
$ hdfs dfs -checksum /project2/file.txt
0000020000000000000000001952d653ccba138f0c4cd4209fbf8e2e
A bit puzzling as I would expect identical checksum to be generated against the identical content.
Checksum for a file can be calculated using hadoop fs command.
Usage: hadoop fs -checksum URI
Returns the checksum information of a file.
Example:
hadoop fs -checksum hdfs://nn1.example.com/file1
hadoop fs -checksum file:///path/in/linux/file1
Refer : Hadoop documentation for more details
So if you want to comapre file1 in both linux and hdfs you can use above utility.
I wrote a library with which you can calculate the checksum of local file, just the way hadoop does it on hdfs files.
So, you can compare the checksum to cross check.
https://github.com/srch07/HDFSChecksumForLocalfile
If you are doing this check via API
import org.apache.hadoop.fs._
import org.apache.hadoop.io._
Option 1: for the value b9fdea463b1ce46fabc2958fc5f7644a
val md5:String = MD5Hash.digest(FileSystem.get(hadoopConfiguration).open(new Path("/project1/file.txt"))).toString
Option 2: for the value 3e50be59553b2ddaf401c575f8df6914
val md5:String = FileSystem.get(hadoopConfiguration).getFileChecksum(new Path("/project1/file.txt"))).toString.split(":")(0)
It does crc check. For each and everyfile it create .crc to make sure there is no corruption.
Given a Parquet dataset distributed on HDFS (metadata file + may .parquet parts), how to correctly merge parts and collect the data onto local file system? dfs -getmerge ... doesn't work - it merges metadata with actual parquet files..
There is a way involving Apache Spark APIs - which provides a solution, but more efficient method without third-party tools may exist.
spark> val parquetData = sqlContext.parquetFile("pathToMultipartParquetHDFS")
spark> parquet.repartition(1).saveAsParquetFile("pathToSinglePartParquetHDFS")
bash> ../bin/hadoop dfs -get pathToSinglePartParquetHDFS localPath
Since Spark 1.4 it's better to use DataFrame::coalesce(1) instead of DataFrame::repartition(1)
you may use pig
A = LOAD '/path/to parquet/files' USING parquet.pig.ParquetLoader as (x,y,z) ;
STORE A INTO 'xyz path' USING PigStorage('|');
You may create Impala table on to it, & then use
impala-shell -e "query" -o <output>
same way you may use Mapreduce as well
You may use parquet tools
java -jar parquet-tools.jar merge source/ target/
I'm looking for efficient way to sync list of directories from one Hadoop filesytem to another with same directory structure.
For example lets say HDFS1 is official source where data is created and once a week we need to copy newly created data under all data-2 directories to HDFS2:
**HDFS1**
hdfs://namenode1:port/repo/area-1/data-1
hdfs://namenode1:port/repo/area-1/data-2
hdfs://namenode1:port/repo/area-1/data-3
hdfs://namenode1:port/repo/area-2/data-1
hdfs://namenode1:port/repo/area-2/data-2
hdfs://namenode1:port/repo/area-3/data-1
**HDFS2** (subset of HDFS1 - only data-2)
hdfs://namenode2:port/repo/area-1/dir2
hdfs://namenode2:port/repo/area-2/dir2
In this case we have 2 directories to sync:
/repo/area-1/data-2
/repo/area-1/data-2
This can be done by:
hadoop distcp hdfs://namenode1:port/repo/area-1/data-2 hdfs://namenode2:port/repo/area-1
hadoop distcp hdfs://namenode1:port/repo/area-2/data-2 hdfs://namenode2:port/repo/area-2
This will run 2 Hadoop jobs, and if number of directories is big, let's say 500 different non overlapping directories under hdfs://namenode1:port/ - this will create 500 Hadoop jobs which is obvious overkill.
Is there a way to inject custom directory list into distcp?
How to make distcp create one job copying all paths in custom list of directories?
Not sure if this answers the problem, but I noticed you haven't used the "update" operator. The "-update" operator will only copy over the difference in the blocks between the two file systems...