[Date Index] [Thread Index] [Author Index]

Re: Re: Intersection of lists of lists based on the first term

• To: mathgroup at smc.vnet.net
• Subject: [mg87533] Re: [mg87492] Re: Intersection of lists of lists based on the first term
• From: "Szabolcs HorvÃt" <szhorvat at gmail.com>
• Date: Sat, 12 Apr 2008 07:00:13 -0400 (EDT)
• References: <ftmu64$50i$1@smc.vnet.net> <200804110956.FAA08360@smc.vnet.net>

On Fri, Apr 11, 2008 at 9:46 PM, Stern <nycstern at gmail.com> wrote:
> Thank you, but I think I am not understanding the second bit -- what is
> (isCommon[#] = True) & /@ timeStampsInCommon supposed to achieve other than
> a list of True equal to the length of timeStampsInCommon in every case?
>

Perhaps I should have written it as

Scan[(isCommon[#] = True) &, timeStampsInCommon]

to emphasize that the function (isCommon[#] = True) & is only invoked
for its side effects, i.e. defining new DownValues for the symbol
isCommon.

This will set the value of isCommon[ ... ] to True for each of the
elements in timeStampsInCommon.  Thus, isCommon[elem] is equivalent to
MemberQ[timeStampsInCommon, elem], but it is much faster if
timeStampsInCommon has many elements (because Mathematica will use a
hash table for looking up elements instead of simply iterating through
timeStampsInCommon).  Note that isCommon[] needs to be Clear[]ed after
each use (a better solution is encapsulating the whole thing into a
Module where isCommon is a local symbol).

• References:
  • Re: Intersection of lists of lists based on the first term
    • From: Szabolcs Horvát <szhorvat@gmail.com>