FileDate in the original suggestion by Daniel only goes back to v3.
To address the poster's original questions: I don't believe that Mathematica has issues in filtering or using a water-fall like method. I can't speak to if Mathematica optimizes such a process.
Hans -----Original Message----- From: Daniel Lichtblau [mailto:email@example.com] Sent: Thursday, May 08, 2014 9:37 AM To: Hans Michel; firstname.lastname@example.org Subject: Re: Chained-functional notation examples?
On 05/08/2014 09:20 AM, Hans Michel wrote: > Daniel: > > Now What's the catch with Today? > > Today will only work in Wolfram Language (Mathematica v 10), which may > be officially released Tomorrow? > > One can go on with a Who's on First theme. (Goes along with W|A > baseball > data) > > But currently the DateObject "Today" would have to be changed to > something that would work in Mathematica 9 or below. > > Hans
Hans et al,
You are correct, I was using an unreleased version. Actually some good came of it (I found a weakness in date comparisons that got addressed). I'll leave modification for versions <=9 to others more adept at this sort of thing. I never was much good with dates...
> -----Original Message----- > From: Daniel Lichtblau [mailto:email@example.com] > Sent: Thursday, May 08, 2014 3:08 AM > To: firstname.lastname@example.org > Subject: Re: Chained-functional notation examples? > > > A query such as this might do what you have in mind. Here > directoryname should be changed to whatever directory you want searched. > > Select[FileNames[("*uid*" | "*UID*"), directoryname, Infinity], > ((StringMatchQ[DirectoryName[#] , "*egal*"] && Today < > DayPlus[FileDate[#], 3]) &] > > Also there is a case sensitive switch if you want to allow strings with e.g. > "uID". If you want to allow that uid only in the name and not full > string > > Daniel Lichtblau > Wolfram Research > > ------------------- > > On Tue, 06 May 2014 06:26:46 +0000, Unknown wrote: > > > I came to 'Mathematica' via Xahlee's criticism of the ad-hoc > nature of > unix-piping [functional notation]. He claims [& I > believe him] that > Mathematica has a better, more consistent > notation. But the facility of > PRE, IN & POST-fix alternatives, > seems bad. You want ONE way of > acieving the goal. More rules just increases mental load. > > > > A major benefit of functional [unix-piping] programming style, is > that > you don't need to remember the-full-journey: you just need to > remember > the previous stage's output. Nor do you need to remember several names: > > the previous output is just "it". A superficial read through a > recent > article[s] here, about <collecting data from several > servers, and > agregating it, and sending the result to a master> > seemed very > interesting, and matches my ideas of using functional > programming. But > I can't afford to invest in ANOTHER > notation/syntax, without good > prospect of productivety increase. > > > > Just as a test, how would Mathematica handle the following [or > part of] >> little task: > > > > search all files in Dir=tree D | > > which are less than N days-old | > > and which contain string S1 | > > and which contain string S2 . > > > > Actually, this seems not a good example, since it's biased towards > the > *nix file system's format/syntax. > > > > Try: > > Search in table of ListOfOpenFiles for lineS with path-P [field] | > > which have same tty-field as line with path-P2 & program-M [field] > > > This sounds like a data-base problem? > > > > Or is there a nice list of 'such' Mathematica examples? > > > > Thanks, > > > > == John Grant. > > Here's a related real-live problem: > list all files in Dir-tree:D | > which are less-than daysOld:N | > and contain "egal" in the FullPathName | > and contain String:"uid" OR "UID" >