Rich: On Thu, Aug 19, 2021 at 6:59 PM Rich Shepard <rshepard@xxxxxxxxxxxxxxx> wrote: > On Thu, 19 Aug 2021, Rob Sargent wrote: > > Yeah, but my quibble is the the table you described up-thread. Your > > contact table contains next_contact? I think that column should be > > normalized out. > Why should I have a separate table with one column: next_contact? The > next_contact date is associated with a person and the date that person was > contacted. Is your next_contact really dependent on the contact record ( person+contact_date? ). I mean, your tables seem CRM like. You stated next_contact for old contacts is not interesting. It seems next_contact is associated just by a person, something like this: You contact person 1 on april 2, decide to follow up on june 2 => insert ( p=1, cd=20210402, nc=20210602 ). Case 1: You contact it in 20210603 , followup for mid july: Insert (1, 20210603, 20210715). Is the nc for the previous record useful for anything ? ( It seems to be nor useful for calling, but you may want to do reports to see next-contact correlation with the contacts on another records ) Case 2: You contact in mid may ( for whatever reason, he calls you, whatever, you insert a record for follow up in July ( 1,20210515, 20210715). Is the next contact date for june still valid? To me, and maybe others, it seems there is only one valid next-contact date, potentially null, which should go in the person record, and may be the contacts table should capture the value of next-contact at the time the contact was made for analysis. This also makes finding contacts to be made soon easier. Of course no separate table for next-contact. Francisco Olarte.
