DynamoDB query() versus getItem() for single-item retrieval based on the index

Question

If I'm retrieving a single item from my table based on the indexed hash key, is there a performance difference between query() or getItem()?

score 57 · Answer 1 · edited Dec 08 '22 at 11:18

57

getItem will be faster

getItem retrieve via hash and range key is a 1:1 fit, the time it takes (hence performance) to retrieve it is limited by the hash and sharding internally.

Query results in a search on "all" range keys. It adds computational work, thus considered slower.

Edit: Just for quick comparison adding the following picture from this blog entry.

edited Dec 08 '22 at 11:18

HFinch

534
1
7
20

answered Sep 03 '12 at 13:01

Chen Harel

9,684
5
44
58

What is a range key? Do you mean sort key? – Sebastian Nielsen Nov 08 '22 at 13:41
@SebastianNielsen back in the day (this answer is from 2012!) DynamoDB called it . Then around the time the document api was released, the conversation shifted to .. – Chen Harel Nov 09 '22 at 15:29

score 31 · Answer 2 · answered Sep 03 '12 at 14:36

31

In Amazon's DynamoDB, your performances are guaranteed whatever the access method. (you pay for it).

There may be a couple a milliseconds differences on the DynamoDB servers themselves as suggested by Chen Harel but these are negligible because of the HTTP request RTT.

This said, it's a good practice to issue a GET instead of QUERY when you have enough informations to do so.

answered Sep 03 '12 at 14:36

yadutaf

6,840
1
37
48

16

Those couple of milliseconds can make a huge difference at scale, even in the context of an HTTP request. That 2-3ms additional latency to DynamoDB could mean 2-3ms time that a request remains open between your web server and your backend service, tying up resources. – Annabelle Jun 26 '15 at 00:04
6

It's good people are thinking of performance. When we talk about optimizations of 2-3ms, there are not many webservices around the world that really need to think of this optimization. 99% of all webservices will never even notice this little delay. – infinity1975 Oct 01 '18 at 07:29

score 26 · Answer 3 · answered Feb 04 '20 at 20:08

26

As suggested by aws employee in one of the discussion, I quote:

The latency of GetItem vs Query with limit=1 will be equivalent.

AWS discussion link

answered Feb 04 '20 at 20:08

Deep Patel

619
7
8

2

Link is outdated – Mr. N Jan 18 '23 at 16:55

score -1 · Answer 4 · answered Feb 21 '19 at 19:53

There is no performance difference between the two. The hash calculation in both the queries are done 1 by 1. The latter, i.e., get item is just provided as an analogy to the JPA repository/spring findOne/findById to make wiring in Spring Bean wiring/ Hibernate configs easier.

DynamoDB query() versus getItem() for single-item retrieval based on the index

4 Answers4

Linked