Re: Re: Re: Bug in Import?

*To*: mathgroup at smc.vnet.net*Subject*: [mg55553] Re: [mg55521] Re: [mg54939] Re: Bug in Import?*From*: "Dale R. Horton" <daleh at wolfram.com>*Date*: Tue, 29 Mar 2005 03:42:35 -0500 (EST)*References*: <cvum3g$bmp$1@smc.vnet.net> <200503030329.WAA21146@smc.vnet.net> <d09c8k$cvk$1@smc.vnet.net> <d0e6pa$osa$1@smc.vnet.net> <200503070710.CAA07704@smc.vnet.net> <200503270743.CAA06265@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

Quite simply, it is a bug, which will be fixed in a future version. We sincerely apologize for any inconvenience. The bug occurs when a field can be parsed as a sequence of numbers. There are multiple workarounds, though. As previously noted, Import["test.tsv", ConversionOptions -> {"Numeric" -> False}] will import all fields as strings. You can then post-process to convert the strings into numbers/dates/etc. For reasons that are complicated, Import["test.tsv", "Table", ConversionOptions -> {"TableSeparators" -> {{"\r", "\n"}, {"\t", "\b"}}} ] will also work. And the elements will get automatically converted. -Dale On Mar 27, 2005, at 1:43 AM, Chris Chiasson wrote: > Was there ever an explanation for the second line being imported > differently? WRI? > > > On Mon, 7 Mar 2005 02:10:23 -0500 (EST), Peter Pein <petsie at arcor.de> > wrote: >> Hans Michel wrote: >> >>> In[1]:= Export["test.tsv", {{"2005-01-24", "2005-02-14"}, >>> {"2005-01-24", \ >>> "2005-02-14"}}] >>> >>> Out[1]= test.tsv >>> >>> In[2]:= Import["test.tsv", ConversionOptions -> {"Numeric" -> False, >>> "DateStyle" \ >>> -> "American", "DateSeparator" -> "-"}] >>> >>> Out[2]= {{2005-01-24, 2005-02-14}, {2005-01-24, 2005-02-14}} >>> >>> In[3]:= Import["test.tsv"] >>> >>> Out[3]= {{2005-01-24, 2005-02-14}, {2005, -1, -24, 2005, -2, -14}} >>> >>> In[4]:= Import["test.tsv", ConversionOptions -> {"NumberSigns" -> >>> {"^", >>> "*"}}] >>> >>> Out[4]= {{2005-01-24, 2005-02-14}, {2005-01-24, 2005-02-14}} >>> >>> In[5]:= $Version >>> >>> Out[5]= 5.1 for Microsoft Windows (October 25, 2004) >>> >>> Window XP Profesional >>> >>> The format for Import is >>> >>> Import["test.tsv", "TSV", Encoding -> None, ConversionOptions -> >>> {...}] >>> >>> That is hard to find in the Help files. >>> >>> Still does not explain why the second line gets messed up. Is it the >>> record >>> separator, is it the number sign >>> >>> Hans >>> >>> No problems in Version 5.0 >>> In[1]:= >>> Export["test.tsv", {{"2005-01-24", "2005-02-14"}, {"2005-01-24", \ >>> "2005-02-14"}}] >>> Out[1]= >>> test.tsv >>> In[2]:= >>> Import["test.tsv", ConversionOptions -> {"Numeric" -> False, >>> "DateStyle" >>> -> "American", "DateSeparator" -> "-"}] >>> Out[2]= >>> {{2005-01-24, 2005-02-14}, {2005-01-24, 2005-02-14}} >>> In[3]:= >>> Import["test.tsv"] >>> Out[3]= >>> {{2005-01-24, 2005-02-14}, {2005-01-24, 2005-02-14}} >>> In[4]:= >>> $Version >>> Out[4]= >>> 5.0 for Microsoft Windows (June 10, 2003) >>> >>> >>> "DrBob" <drbob at bigfoot.com> wrote in message >>> news:d09c8k$cvk$1 at smc.vnet.net... >>> >>>> Not well indexed? That's putting it mildly. >>>> >>>> Here's an example of use (MUCH needed in Help): >>>> >>>> Export["test.tsv", >>>> {{"2005-01-24", "2005-02-14"}, >>>> {"2005-01-24", "2005-02-14"}}] >>>> Import["test.tsv", >>>> ConversionOptions -> {"Numeric" -> >>>> False, "DateStyle" -> "American", "DateSeparator" -> "-"}] >>>> >>>> No matter what options are used, however... >>>> >>>> Why is the second record converted differently than the first? >>>> >>>> Bobby >>>> >> [...] >>>> -- >>>> DrBob at bigfoot.com >>>> www.eclecticdreams.net >>>> >>> >>> >> I think, Bobby wanted to say: "If there's a bug, why only in the >> second >> line?" >> It seems a "buggy bug" to me too. >> >> -- >> Peter Pein >> Berlin >> >> > > > -- > Chris Chiasson > Kettering University > Mechanical Engineering > Graduate Student > 1 810 265 3161 >

**References**:**Re: Bug in Import?***From:*"Hans Michel" <hmichel@sdc.cox.net>

**Re: Bug in Import?***From:*Peter Pein <petsie@arcor.de>

**Re: Re: Bug in Import?***From:*Chris Chiasson <chris.chiasson@gmail.com>