A shortcut through long loops: An illustration of two alternatives to looping over observations

It is well known that looping over observations can be slow and should be avoided. The objective of this article is to discuss two alternative solutions to looping over observations that can be used to overcome a particular data-management problem of merging datasets in which unique key identifiers changed over time. The first alternative, mapch, which is introduced in this article, uses a combination of appending, indexing, and merging to solve the problem, while the second alternative uses repeated merging. Both solutions are much quicker than looping over observations. However, depending on the nature of the problem, one solution may work better than the other. It is argued that the use of such dataset-type manipulations may be suitable to overcome other data-management problems. More generally speaking, the issue that is addressed—searching for an alternative to looping over observations—may be common and illustrates the importance of balancing the costs of developing an efficient solution with the benefits accruing from that solution.


Issue Date:
2008
Publication Type:
Journal Article
DOI and Other Identifiers:
dm0041 (Other)
PURL Identifier:
http://purl.umn.edu/122620
Published in:
Stata Journal, Volume 08, Number 4
Page range:
540-553
Total Pages:
14

Record appears in:



 Record created 2017-04-01, last modified 2017-08-26

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)