R dataframe as in-memory database

Received: by 10.151.108.5 with SMTP id k5mr502041ybm.8.1229971521560; Mon, 22
Dec 2008 10:45:21 -0800 (PST)
Date: Mon, 22 Dec 2008 10:45:21 -0800 (PST)
X-IP: 129.170.66.129
User-Agent: G2/1.0
X-Google-Token: Zq9xmAwAAAD663sZREJznzBJUhYCYZcj
X-HTTP-UserAgent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US;
rv:1.9.0.5) Gecko/2008120121 Firefox/3.0.5,gzip(gfe),gzip(gfe)
Message-ID: <647fbf98-e633-4d27-b4cb-e510cf19c0be@q18g2000vbn.googlegroups.com>
Subject: R dataframe as in-memory database
From: braver
To: “Kdb+ Personal Developers”
X-Google-Approved: charlie@kx.com via web at 2008-12-22 18:56:11

It occurred to me that Q’s in-memory database is in fact similar to an
R dataframe, which – each dataframe – is its own in-memory
database. With R, you can select by logical conjunctions:

z[z$person_id==i && in_range(z$time,start,finish),]

What’s other folks impression on R vs Q comparison? Both can do
vector operations as well. With 64-bit R, which is finally reliably
here, you can read a whole series of tables, an entire database, into
memory as a series of dataframes, and execute fast queries on them by
logical selection.

Cheers,
Alexy