External Sort  External Sort

home page up Google search web for more information on this topic jump to foot of page translate this page with Babelfish ©1996-2012 Roedy Green, Canadian Mind Products

This essay does not describe an existing computer program, just one that should exist. This essay is about a suggested student project in Java programming. This essay gives a rough overview of how it might work. I have no source, object, specifications, file layouts or anything else useful to implementing this project.

This project outline is not like the artificial, tidy little problems you are spoon-fed in school, when all the facts you need are included, nothing extraneous is mentioned, the answer is fully specified, along with hints to nudge you toward a single expected canonical solution. This project is much more like the real world of messy problems where it is up to you to fully the define the end point, or a series of ever more difficult versions of this project, and research the information yourself to solve them.

Everything I have to say to help you with this project is written below. I am not prepared to help you implement it; or give you any additional materials. I have too many other projects of my own.

Though I am a programmer, I don’t do people’s homework for them. That just robs them of an education.

You have my full permission to implement this project in any way you please and to keep all the profits from your endeavour.

Please do not email me about this project without reading the disclaimer above.

What do you do when a sort is too big to fit in RAM (Random Access Memory)? When Sun’s Collections. sort fails? You need some sort of external sort that uses intermediate disk space. I can think of four ways to implement this:

  1. Create an easy-to-use interface to an existing external sort such as the Unix sort or Opttech sort.
  2. Create a traditional external sort that sorts records made of an array of bytes, the way mainframes have done it since the 1960s. The problem is the user must convert his data to byte arrays with DataOutputStream, and specify sort keys in terms of inconvenient byte offsets, then reconstitute Objects with a DataInputStream from the bytes when the sort is done.
  3. Implement a tag sort. You create small records containing only the key and relative record number, and sort them with an internal sort. They might just fit, when the entire dataset would not.
  4. Create a sort that behaves just like Sun’s that transparently uses extra disk space when the data set is large. The advantage is the user can specify the sort with a Comparator, using Java field names and operators. The disadvantage is the intermediate files would have to be stored as serialised records. There is considerable serialization overhead, both in terms of speed and size. Further the records to be sorted have to be serializable, something not required for an internal sort. Unfortunately, to compare records, the intermediate records must be reconstituted at every stage of the sort. Even though this would be quite slow, it might be the most popular version since it could be slid in without changing the application. To convert to a byte-array sort would be a lot of work, to a primitive addressing scheme. That effort might better be spend converting to an SQL (Standard Query Language) database that has external sorting built-in.
An external sort typically sorts a chunk of N records in RAM then writes it out to a file. It repeats, each time writing it out to a different file. When it has written K files (where optimum N and K are to be determined by experiment. K will likely turn out be in the range 4…30), it reuses the files, appending to the end.

Then it does a K-way merge, reading from K files and writing the merged results of the first chunk in each file to another file. You read the files with a buffered reader with buffer size B, (where again the optimum size of B is determined by experiment.) Then it merges the second chunks etc. Each merge pass reduces the number of chunks to 1/K, and makes each chunk K times as large. You repeat until all the chunk are merged into one sorted one.

How cleverly you estimate values for N, K and B will largely determine the performance of your sort.

external sorting
HTML table sorter
JASM
NIST on external sorts
on the fly compose, compile and run
sort comparator
sort visualiser
CMP homejump to top You can get the freshest copy of this page from: or possibly from your local J: drive (Java virtual drive/mindprod.com website mirror)
http://mindprod.com/project/externalsort.html J:\mindprod\project\externalsort.html
logo
Please email your , letters to the editor, errors, omissions, typos, formatting errors, ambiguities, unclear wording, broken/redirected link reports, suggestions to improve this page or comments to Roedy Green : feedback email. If you want your message, your name or email kept confidential, not considered for public posting, please explicitly specify that. Unless you state otherwise, I will treat your message as a letter to the editor that I may or may not publish in the feedback section. After that, it will be too late to retract it. If you disagree with something I said, please quote it and cite the web page where you found it, tell me why you think it is wrong, and, if possible, provide some supporting evidence. Threatening to kill me or spouting obscenities has yet to persuade me to change my mind.
mindprod.com IP:[65.110.21.43]
view BlogYour face IP:[38.107.179.212]
You are visitor number 11.