fgetcsv not reading enclosures

问题

I am trying to read a CSV line using fgetcsv. However it seem not to care at all about enclosures.

Here is what the line looks like:

Super Administrator,"ROLE_SUPER_ADMIN, ROLE_GROUP_GUEST, ROLE_GROUP_WRITER, ROLE_USER_WRITER, ROLE_USER_GUEST"

This is a valid line by CSV standards, and it should return the following with " as the enclosure and , as the delimiter:

$l = fgetcsv($handle, 0, ',', '"');
array(
    [0] => 'Super Administrator',
    [1] => 'ROLE_SUPER_ADMIN, ROLE_GROUP_GUEST, ROLE_GROUP_WRITER, ROLE_USER_WRITER, ROLE_USER_GUEST',
);

However this is what I get:

$l = fgetcsv($handle, 0, ',', '"');
array(
    [0] => 'Super Administrator',
    [1] => '"ROLE_SUPER_ADMIN',
    [2] => ' ROLE_GROUP_GUEST',
    [3] => ' ROLE_GROUP_WRITER',
    [4] => ' ROLE_USER_WRITER',
    [5] => ' ROLE_USER_GUEST"',
);

So yeah, it seems fgetcsv is right-out ignoring completely the enclosure character. I tried to get it working using fgets and then str_getcsv but the result is the same - str_getcsv is most probably called by fgetcsv for the CSV conversion anyway.

I figured it could be some sort of regression bug that found its way in the PHP version I'm using, so I'm pasting the result of a php -v command here:

me@linux:~/$ php -v
PHP 5.5.27-1+deb.sury.org~trusty+1 (cli) (built: Jul 15 2015 12:14:44) 
Copyright (c) 1997-2015 The PHP Group
Zend Engine v2.5.0, Copyright (c) 1998-2015 Zend Technologies
    with Zend OPcache v7.0.6-dev, Copyright (c) 1999-2015, by Zend Technologies

I believe this is the latest version of PHP5 at the moment I'm writing this. Although I suspect this is some kind of PHP bug, it could also be some installation/configuration problem, hence why I'm publishing this question.

Has anyone already experienced this? What is the best way to solve this problem?

EDIT:
As pointed out by Mark Baker, PHP5.5.27-1 is actually the latest revision of PHP5.5 at the time being.

EDIT 2:
Here is an attempt with the bin2hex function, displayed with a var_dump.

Line read ",","," (2 cells containing only a coma in each cell): 0022002c0022002c0022002c0022000a.

With bin2hex('"') the result is just 22.

str_getcsv still gives me this error though.

回答1:

OK, solved.

This is what everyone was suspecting: the encoding of the file was messed up. I could not know which encoding this was, but LibreOffice proposed me Unicode whenever I tried to open the CSVs.

I had to open them with nano to realize there was indeed an encoding problem. Gedit, vim or any other tool I had on my computer raised no errors. When opened with nano, an @ symbol was inserted between every other characters and line feeds were not read correctly.

It seems there are some encodings that are not well supported by fgetcsv. To solve the problem, I recreated the files from nano (copy-paste from another tool that did not display the @).

来源：https://stackoverflow.com/questions/31860061/fgetcsv-not-reading-enclosures

标签

php

csv

fgetcsv

php-5.5