U++ forum: Welcome to the forum

Re: Quick explanation of function calls for postgresql [message #16613 is a reply to message #16598]

Wed, 02 July 2008 13:01

mirek
Messages: 13976
Registered: November 2005

Ultimate Member

captainc wrote on Mon, 30 June 2008 16:50

I've been looking over the examples for postgres as I would like to use the functionality, but I am confused as to what exactly some of the functions do. I have been browsing the source, but I think it would help if we had a quick explanation of the structure of the code and some of these functions.

from the example, this is the section that confuses me most:

Progress p;
p.SetText(t_("Creating _DEBUG database"));
SqlSchema sch(PGSQL);
StdStatementExecutor se(m_session);
All_Tables(sch);
if(sch.ScriptChanged(SqlSchema::UPGRADE))
PostgreSQLPerformScript(sch.Upgrade(),se, p);
if(sch.ScriptChanged(SqlSchema::ATTRIBUTES)) {
PostgreSQLPerformScript(sch.Attributes(),se, p);
}
if(sch.ScriptChanged(SqlSchema::CONFIG)) {
PostgreSQLPerformScript(sch.ConfigDrop(),se, p);
PostgreSQLPerformScript(sch.Config(),se, p);
}
sch.SaveNormal();

All these are related to "uploading" .sch file into database.

SqlSchema is instance to hold uploading SQL scripts.

All_Tables will "dump" SQL scripts from .sch file and stores them into database. This function is defined by including .sch file. It is sort of .sch interface point.

There is a couple of "standard" scripts.

UPGRADE script contains creates all tables and columns, however does so by creating tables only with the first column, then adding all columns (using ALTER TABLE ... ADD COLUMN) one by one. This allows incemental development of the modle - at start of app, model gets upgraded, commands to create all columns and tables that already exist simply fail, but any new columns or tables are added. Note that this model does not support removing columns or changing datatype - that has to be done manually.

ATTRIBUTES script adds any "attributes", namely constraints or indicies. The main reason to have this separated is because sometimes contraints have to be added only after the tables are defined (forward foreign keys). Secondary reason is that U++ also generates "drop" scripts; sometimes it is useful, when maintaining app, to drop all or some contraints and indicies and recreate them later.

CONFIG contains any data configuration - usually inserts of "metadata". Frankly, it is less important and in many cases this can be easily done some other way.

SaveNormal saves all scripts into U++ config dir (.exe dir in Win32, ~/.upp/appname in Linux).

sch.ScriptChanged compares script stored in sch with script stored by SaveNormal in previous run of code. This is to avoid running scripts each time (it can be time consuming). OTOH, sometimes this incorrectly blocks script execution, for example, if you are for some reason using two identical database schemas, you upload to one (thus in this session SaveNormal overwrites script files), then run app again with second schema - and files are identical. If I have this problem, I usually just delete scripts (they are not equal then Wink

.

Mirek

[Updated on: Wed, 02 July 2008 13:05]

Report message to a moderator