Schedhistory in the database¶

Table

Once we know the SID of the schedule we would like to assign to a given employee, it is time to insert a record into the "schedhistory" table:

      CREATE TABLE IF NOT EXISTS schedhistory (
        shid       serial PRIMARY KEY,
        eid        integer REFERENCES employees (eid) NOT NULL,
        sid        integer REFERENCES schedules (sid) NOT NULL,
        effective  timestamp NOT NULL,
        remark     text,
        stamp      json
      );

Stored procedures

This table also includes two stored procedures -- "sid_at_timestamp" and "current_schedule" -- which will return an employee's schedule as of a given date/time and as of 'now', respectively. For the procedure definitions, see "dbinit_Config.pm"

See also "When history changes take effect".

Schedhistory in the Perl API¶

• constructor (spawn)
• reset method (recycles an existing object)
• basic accessors (shid, eid, sid, effective, remark)
• load_by_eid method (load schedhistory record from EID and optional timestamp)
• load_by_shid method (wrapper for load_by_id)
• load_by_id (load schedhistory record by its SHID)
• insert method (straightforward)
• delete method (straightforward) -- not tested yet # FIXME

For basic workflow, see "t/model/schedule.t".

load_by_eid¶

Class method. Given an EID, and, optionally, a timestamp, attempt to look it up in the database. Generate a status object: if a schedhistory record is found, it will be in the payload and the code will be 'DISPATCH_RECORDS_FOUND'.

load_by_id¶

Given a shid, load a single schedhistory record.

load_by_shid¶

Wrapper for load_by_id

insert¶

Instance method. Attempts to INSERT a record into the 'Schedhistory' table. Field values are taken from the object. Returns a status object.

update¶

Instance method. Updates the record. Returns status object.

delete¶

Instance method. Deletes the record. Returns status object.

get_schedhistory¶

Takes a PARAMHASH which can have one or more of the properties 'eid', 'nick', and 'tsrange'.

At least one of { 'eid', 'nick' } must be specified. If both are specified, the employee is determined according to 'eid'.

The function returns the history of schedule changes for that employee over the given tsrange, or the entire history if no tsrange is supplied.

The return value will always be an App::CELL::Status object.

Upon success, the payload will be a reference to an array of "schedhistory" objects. If nothing is found, the array will be empty. If there is a DBI error, the payload will be undefined.

Sam Wallace joins the firm¶

Let's say Sam's initial schedule is 09:00-17:00, Monday to Friday. To reflect that, the "schedintvls" table might contain the following intervals for "sid = 9"

    '[2014-06-02 09:00, 2014-06-02 17:00)'
    '[2014-06-03 09:00, 2014-06-03 17:00)'
    '[2014-06-04 09:00, 2014-06-04 17:00)'
    '[2014-06-05 09:00, 2014-06-05 17:00)'
    '[2014-06-06 09:00, 2014-06-06 17:00)'

and the "schedhistory" table would contain a record like this:

    shid      848 (automatically assigned by PostgreSQL)
    eid       39 (Sam's Dochazka EID)
    sid       9
    effective '2014-06-04 00:00'

(This is a straightfoward example.)

Sam goes on night shift¶

A few months later, Sam gets assigned to the night shift. A new "schedhistory" record is added:

    shid     1215 (automatically assigned by PostgreSQL)
    eid        39 (Sam's Dochazka EID)
    sid        17 (link to Sam's new weekly work schedule)
    effective  '2014-11-17 12:00'

And the schedule intervals for "sid = 17" could be:

    '[2014-06-02 23:00, 2014-06-03 07:00)'
    '[2014-06-03 23:00, 2014-06-04 07:00)'
    '[2014-06-04 23:00, 2014-06-05 07:00)'
    '[2014-06-05 23:00, 2014-06-06 07:00)'
    '[2014-06-06 23:00, 2014-06-07 07:00)'

(Remember: the date part in this case designates the day of the week)