Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!gegeweb.org!de-l.enfer-du-nord.net!feeder1.enfer-du-nord.net!tudelft.nl!txtfeed1.tudelft.nl!feed.xsnews.nl!border-1.ams.xsnews.nl!newspump.sol.net!posts.news.twtelecom.net!nnrp2.twtelecom.net!not-for-mail From: Frank Iannarilli Newsgroups: comp.soft-sys.math.mathematica Subject: Re: Efficient way to read binary data line-by-line Date: Fri, 30 Dec 2011 12:14:47 +0000 (UTC) Organization: Steven M. Christensen and Associates, Inc and MathTensor, Inc. Sender: steve@smc.vnet.net Approved: Steven M. Christensen , Moderator Message-ID: References: <200909090846.EAA06073@smc.vnet.net> Reply-To: comp.soft-sys.math.mathematica@googlegroups.com Lines: 6 NNTP-Posting-Date: 30 Dec 2011 11:19:52 GMT NNTP-Posting-Host: 4a547f8c.news.twtelecom.net X-Trace: DXC=2Q5K`Z?A276BdhbaOf?SM9C_A=>8kQj6=hHXa^^g6TZ4H7^L0=Y;R<>EFiONJ7[Go6jeVP\D4VId7 X-Complaints-To: abuse@twtelecom.net Xref: x330-a1.tempe.blueboxinc.net comp.soft-sys.math.mathematica:7482 Leonid is spot-on with the sequential nature of BinaryReadList. But BinaryRead is EXTREMELY inefficient compared to BinaryReadList (at least for version 8.0.1). I just took a look at why, and it is apparently because the result from BinaryRead is NOT in PackedArray format (per `Developer`PackedArrayQ), even for a 65K-sized read request. OTOH, BinaryReadList returns a payload as a PackedArray, which implies its conversions from Bytes to requested type will be MUCH faster.