Issue488MostExportsDontIncludeStudyTime

Merged
#125 · Created  · Last updated

Merged pull request

Merged in Issue488MostExportsDontIncludeStudyTime (pull request #125)

e1dc2ae·Author: ·Closed by: ·2017-11-17

Description

  • Refs #488, #556, #484

  • Ready for review. Implements consistent common columns across all modalities, csv and xlsx.

  • Massively improved code, removal of duplication. Could definitely be improved further.

  • Improved efficiency too. Timings:

Export type

number of records

0.7.4

this branch

CT XLSX

183

16 minutes, 24 seconds

7 minutes, 12 seconds

CT CSV

183

6 minutes, 56 seconds

6 minutes, 40 seconds

RF XLSX

22

1 hour 13 minutes

1 hour 1 minute

RF CSV

22

22.4 seconds

15.8 seconds

RF CSV

47

47.2 seconds

31.8 seconds

DX XLSX

59

4 minutes 18 seconds

2 minutes 4 seconds

DX XLSX

205

15 minutes 35 seconds

6 minutes 39 seconds

DX CSV

205

5 minutes 36 seconds

5 minutes 50 seconds

MG CSV

137

14 minutes 37 seconds

8 minutes 16 seconds

MG XLSX

137

 

8 minutes 47 seconds

Timings are from identical virtual machines running on a server with lots of capacity. Database is the same on both (dump from production database), data is last x days of a variety of different manufacturers and models.

One of the consequences of the better code is that the XLSX exports are now usually not much slower than CSV exports. And for the extra short while, you get more organisation and summarising, plus no issues with non-ASCII data. This is still a problem when opening CSV in Excel as it refuses to default to assuming the UFT-8 encoding has been used.

0 attachments

0 comments

Loading commits...