I'm using str_getcsv to parse tab separated values being returned from a nosql query however I'm running into a problem and the only solution I've found is illogical.
Here's some sample code to demonstrate (FYI, it seems the tabs aren't being preserved when showing here)...
$data = '0 16 Gruesome Public Executions In North Korea - 80 Killed http://www.youtube.com/watch?v=Dtx30AQpcjw&feature=youtube_gdata "North Korea staged gruesome public executions of 80 people this month, some for offenses as minor as watching South Korean entertainment videos or being fou... 1384357511 http://gdata.youtube.com/feeds/api/videos/Dtx30AQpcjw 0 The Young Turks 1 2013-11-13 12:53:31 9ab8f5607183ed258f4f98bb80f947b4 35afc4001e1a50fb463dac32de1d19e7';
$data = str_getcsv($data,"\t",NULL);
echo '<pre>'.print_r($data,TRUE).'</pre>';
Pay particular attention to the fact that one column (beginning with "North Korea...." actually starts with a double quote " but doesn't finish with one. This is why I supply NULL as the third parameter (enclosure) to override the defaut " enclosure value.
Here is the result:
Array
(
[0] => 0
[1] => 16
[2] => Gruesome Public Executions In North Korea - 80 Killed
[3] => http://www.youtube.com/watch?v=Dtx30AQpcjw&feature=youtube_gdata
[4] =>
[5] => North Korea staged gruesome public executions of 80 people this month, some for offenses as minor as watching South Korean entertainment videos or being fou... 1384357511 http://gdata.youtube.com/feeds/api/videos/Dtx30AQpcjw 0 The Young Turks 1 2013-11-13 12:53:31 9ab8f5607183ed258f4f98bb80f947b4 35afc4001e1a50fb463dac32de1d19e7
)
As you can see the quote is breaking the function. Logically I thought I would be able to use NULL or and empty string'' as the third parameter for str_getcsv (enclosure) but neither worked?!?!
The only thing I could use to get str_getcsv to work properly was a space char ' '. That doesn't make any sense to me becuase none of the columns have whitespace starting and/or ending them.
$data = '0 16 Gruesome Public Executions In North Korea - 80 Killed http://www.youtube.com/watch?v=Dtx30AQpcjw&feature=youtube_gdata "North Korea staged gruesome public executions of 80 people this month, some for offenses as minor as watching South Korean entertainment videos or being fou... 1384357511 http://gdata.youtube.com/feeds/api/videos/Dtx30AQpcjw 0 The Young Turks 1 2013-11-13 12:53:31 9ab8f5607183ed258f4f98bb80f947b4 35afc4001e1a50fb463dac32de1d19e7';
$data = str_getcsv($data,"\t",' ');
echo '<pre>'.print_r($data,TRUE).'</pre>';
Now the result is:
Array
(
[0] => 0
[1] => 16
[2] => Gruesome Public Executions In North Korea - 80 Killed
[3] => http://www.youtube.com/watch?v=Dtx30AQpcjw&feature=youtube_gdata
[4] =>
[5] => "North Korea staged gruesome public executions of 80 people this month, some for offenses as minor as watching South Korean entertainment videos or being fou...
[6] => 1384357511
[7] => http://gdata.youtube.com/feeds/api/videos/Dtx30AQpcjw
[8] => 0
[9] => The Young Turks
[10] =>
[11] =>
[12] =>
[13] =>
[14] => 1
[15] => 2013-11-13 12:53:31
[16] => 9ab8f5607183ed258f4f98bb80f947b4
[17] => 35afc4001e1a50fb463dac32de1d19e7
)
So my question is, why does it work with a space as the enclosure, but not NULL or and empty string? Also are there repercussions to this?
UPDATE 1: It seems this reduced the number of errors I was receiving in our logs but it didn't eliminate them, so I'm guessing that the I used as the enclosure has caused unintended side effects, albeit less troubling than the previous problem. But my question remains the same, why can't I use NULL, or an empty space as the enclosure, and secondly, is there a better way of dealing with / doing this?
Just to give a starting point ...
You might wanna consider working with the string itself, instead of using a function like str_getcsv in your case.
But be aware that there are at least some pitfalls, if you choose this route (might be your only option though):
If you know that you don't have any other TABS in your string other than those ending the fields, and you don't have any linebreaks other than those delimiting a row, you might be fine with this:
$data = explode("\n", $the_whole_csv_string_block);
foreach ($data as $line)
{
$arr = explode("\t", $line);
// $arr[0] will have every first field of every row, $arr[1] the 2nd, ...
// Usually this is what I want when working with a csv file
// But if you rather want a multidimensional array, you can simply add
// $arr to a different array and after this loop you are good to go.
}
Otherwise this is just a starting point for you, to begin and tweak it to your individual situation, hope it helps.