Skip site navigation (1) Skip section navigation (2)

Re: Bug in 7.4_213 driver: returns VARCHAR instead of

From: Oliver Jowett <oliver(at)opencloud(dot)com>
To: "j(dot)random(dot)programmer" <javadesigner(at)yahoo(dot)com>
Cc: Kris Jurka <books(at)ejurka(dot)com>, pgsql-jdbc(at)postgresql(dot)org
Subject: Re: Bug in 7.4_213 driver: returns VARCHAR instead of
Date: 2004-07-30 00:03:25
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgsql-jdbc
j.random.programmer wrote:

> Why are the getXXXStream methods "rarely the correct
> thing" ?
> Ignoring current implementation artifacts (if any),
> they
> are almost always the _right_ thing for large amounts
> of
> data.

The point is that it *is* all about current implementation artifacts. 
There's no way to stream a text value from the server on demand using 
the current protocol. So if you use getXXXStream() with the current 
driver, you always end up wrapping data that's completely on-heap anyway 
with an unnecessary stream interface. Note that this is more of a 
client/server protocol issue than a driver issue (and it's not obvious 
how to solve it at the protocol level anyway).

As Barry said, returning LONGVARCHAR implies the preferred access method 
is via getXXXStream() (see the jdbc type mapping tables), but those 
aren't the best methods for accessing text columns in our current 
implementation. And the metadata should reflect the characteristics of 
the driver implementation actually used, surely..


In response to

pgsql-jdbc by date

Next:From: Nick SelvaDate: 2004-07-30 04:18:23
Subject: Calling PGSQL Stored function thru JDBC
Previous:From: j.random.programmerDate: 2004-07-29 22:57:20
Subject: Re: Bug in 7.4_213 driver: returns VARCHAR instead of

Privacy Policy | About PostgreSQL
Copyright © 1996-2018 The PostgreSQL Global Development Group