[Gnumed-devel] Backend support for a record of incoming data "jobs"

[Jim Busser]

When looked-for results cannot be found in an EMR, the questions are

- did the patient / specimens make it to the lab?
- was there any problem or delay in the lab?
- is the result available?
- did we receive it yet?

An easy "check" would be

        when did the EMR last receive and process *any* results.

For that reason (and more) I would like to suggest a table

        incoming_data_log

with fields

        job_start                       datetime
        job_end                 datetime (or "duration" of type interval)
        process_ip
        process_num
        process_user
        db_user
        data_source_ip
        data_source_url
        file_specification
        script_specification
        encounter_number
        records_imported
        exit_status
        exit_comments

what I was thinking in regard to "process_" fields was to identify the machine 
from which the download / import was being driven… thinking these could help IT 
support to track down problems in case the processes had problems associated.

Not all fields would be used for every transaction:

- the table might separately capture automated download and import events
- downloaded files might, in some cases, be obtained by a manual step via web 
login; in such cases, there would be no record of data sources
- imports would reference the file specification of whatever the user had 
manually deposited into the directory watched by the importer (cron) script
- records of download-only would not populate the encounter_number and 
records_imported fields

-- Jim