====Scattered vs Synchronised Data Sources====

There are two ways that dates and times tend to be stored in databases.

They can be **synchronised**, meaning that every measurement is performed at the same effective time - for example, every point will be recorded every 10 seconds, and each point in that sample pass will have the same timestamp.

^Name^Time^Sample^
|Channel A|4:00:00|12|
|Channel A|4:00:10|12.2|
|Channel A|4:00:20|12.4|
|Channel B|4:00:00|4|
|Channel B|4:00:10|5|
|Channel B|4:00:20|6|
|Channel C|4:00:00|3|
|Channel C|4:00:10|2.8|
|Channel C|4:00:20|2.6|

Or they can be **scattered**, meaning that samples are taken at different times and time stamps are quite randomised. 

^Name^Time^Sample^
|Channel A|3:59:12|12|
|Channel B|4:00:00|4|
|Channel B|4:00:30|6|
|Channel C|4:00:10|3|
|Channel C|4:00:15|2.7||

Database drivers support both kind of data source. However, //synchronised// data sources query significantly faster than //scattered// ones, because it is much more efficient to look up the previous sample time for **one** channel than it is to find the previous sample time for **thousands**.