Search API - Syndication date

[Gordon Paynter]

Hi:

Is the syndication_date in the DigitalNZ API still defined as: "the
date the record was added to DigitalNZ" as per
http://www.digitalnz.org/api-docs/search-records-api ?

I ask because it appears to be returning "old" records? By which I
mean records of things that I would have assumed were added to
DigitalNZ years ago (e.g. Maclean Papers images). However, it may be
that these really are the most recently added things?

Example:
http://api.digitalnz.org/records/v2.json?api_key=XXX&search_text=*:*&num_results=100&sort=syndication_date&direction=desc

Note the purpose of syndication_date was always to get the "freshest"
material from DigitalNZ. If this has changed, is there a way to sort
and get the most-recently-added records? For example, search and sort
by id descending?

Thanks,
Gordon

[Chris McDowall]

Hi Gordon,

Thanks for your question. I need to double-check a couple of things to
reply with certainty. I will get back to you later today.

Cheers,
Chris

On Dec 16, 12:40 am, Gordon Paynter <gordon.payn...@gmail.com> wrote:
> Hi:
>
> Is the syndication_date in the DigitalNZ API still defined as: "the
> date the record was added to DigitalNZ" as perhttp://www.digitalnz.org/api-docs/search-records-api?
>
> I ask because it appears to be returning "old" records? By which I
> mean records of things that I would have assumed were added to
> DigitalNZ years ago (e.g. Maclean Papers images).  However, it may be
> that these really are the most recently added things?
>

> Example:http://api.digitalnz.org/records/v2.json?api_key=XXX&search_text=*:*&...

[Chris McDowall]

Hello Gordon,

We about this in person, but I just realised I never updated the
group.

I investigated the issue and it seems that syndication_date is not
working as intended. Instead of the behaviour you identified above, it
is actually being reset every time an item is re-indexed. This is
(generally) not a problem for collections that we harvest from feeds,
but it is an issue for collections where all items in that collection
are periodically re-harvested in bulk (this is the case for some
National Library collections).

I have written the issue up and we will correct it over the coming
weeks.

Chris

On Dec 16 2011, 8:14 am, Chris McDowall <chris.mcdow...@gmail.com>
wrote: