Re: AW: DB2 terminology

On 21 Aug 2013 14:15, Vern Hamberg wrote:

CREATE SCHEMA also lets you specify lots of other stuff, like tables
and <<SNIP>>

Yes. Like the examples I gave that I suggested would need to be tested, to determine if the documentation is /incorrect/ in its implication. The first request includes a CREATE TABLE as part of the CREATE SCHEMA [described as /schema processor/ syntax in v5r1 docs]. But is that syntax allowed in CREATE COLLECTION, as effectively is alluded in the more recent docs? That is, an implication in the newer docs is that the keyword COLLECTION is a /syntax alternative/ for the keyword SCHEMA. However that is *not* true at v5r3, but has been implied as such, since v5r2; the release at which only CREATE SCHEMA is the documented DDL for creating a library, with some /notes/ about the COLLECTION keyword [and possibly also the DATABASE keyword] as synonyms.

There are also options about data dictionary

Deprecated. They should be avoided except if using a product that depends on them, and the limitations imposed for data types and whatever else are acceptable.

and other things. There might even be an option not to create the
journal stuff and the local copy of the catalog files. <<SNIP>>

AFaIK that still is unchanged; nothing optional, including no ability for CREATE to /overlook/ that the library name already exists. However the QSQXRLF [undocumented] API could be used to remove the SQL Catalog VIEW files, but the SQL will AFaIK still create some of the catalog VIEWs automatically when certain objects are created into the SCHEMA; i.e. for those VIEWs added at\since v2r1m1 IIRC. That API could also be used, in conjunction with STRJRNLIB, to make an effective SCHEMA without any explicitly requested SQL statement; i.e. CRTLIB instead. Even the WITH DATA DICTIONARY clause will still create a *DTADCT object, even though likely still almost no test-cases exist and the horrible problems they can cause if the DtaDct exists presumably still exist; the type and number of restrictions may not be sufficient to have thoroughly discouraged their use.

Regards, Chuck

On 21 Aug 2013 10:46, CRPence wrote:

On 21 Aug 2013 04:12, rob@xxxxxxxxx wrote:

CREATE COLLECTION THING1
CREATE SCHEMA THING2
both create a journal and a receiver and all the same views over
the system catalog objects.<<SNIP>>

More telling, to distinguish each and a possible error\deficiency
in the documentation, would be the difference in effects between
each of:

CREATE COLLECTION THING1 CREATE TABLE LURCH (ITT CHARACTER)

CREATE SCHEMA THING2 CREATE TABLE LURCH (ITT CHARACTER)

CREATE COLLECTION ADDAMS_FAMILY1 FOR ADDAMS1

CREATE SCHEMA ADDAMS_FAMILY2 FOR ADDAMS2