PHP和CSV：删除或忽略以不同长度行的字符开头的行

Question

I have CSV file that uses Keys in the first column, and each has a different row length.

Top of the file, the header row starts with 'M', and then rows starting with 'C' and then 'A' alternate throughout the file like this.

M   P395, 177   177, 13/03/13, , , , , , , FALSE,   1904.2, , , , , , , , , , , , , , 
C   QTM0039326, X6  100013424,  Example, , Example  WA  6754    AU, FALSE,  TRUE    FALSE, N,   FALSE, FALSE, FALSE Example Example Brisbane,   Brisbane City   QLD 4000    AU, , , , , , Example   TRACKADV
A   0.1, , , FALSE  FALSE   0, , , , , , , , , , , , , , , , , , , , , 
C   QTM0039226  7021130 X6  100013427,  Example, , Example  NSW 2795    AU  427181931   FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV
A   4.1 0   0   0, ARTICLE CONTAINS CONSUMER GOOD(S)    FALSE   FALSE   0   0,  FALSE   FALSE   FALSE   FALSE   FALSE, FALSE, , , , , , , , , , , , , , , , 
C   QTM0039214  7021130 X6  100013440   Example, Example, , Example QLD 4502    AU  32858429    FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV
A   1.35    0   0   0, ARTICLE CONTAINS CONSUMER GOOD(S)    FALSE   FALSE   0   0,  FALSE   FALSE   FALSE   FALSE   FALSE, FALSE, , , , , , , , , , , , , , , , 
C   QTM0039296  7021130 X6  100013349, Metro Auto Spares    Example, , Example  TAS 7310    AU  427236691   FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV
A   5.25    0   0   0, ARTICLE CONTAINS CONSUMER GOOD(S)    FALSE   FALSE   0   0,  FALSE   FALSE   FALSE   FALSE   FALSE, FALSE, , , , , , , , , , , , , , , , 
C   QTM0039300  7021130 X6  100013345,  Example, , Example  QLD 4303    AU  402131430   FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV
A   0.6 0   0   0, ARTICLE CONTAINS CONSUMER GOOD(S)    FALSE   FALSE   0   0,  FALSE   FALSE   FALSE   FALSE   FALSE, FALSE, , , , , , , , , , , , , , , , 
C   QTM0039242  7021130 X6  100008683,  Example, , Example  SA  5034    AU  403468706   FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV
A   0.6 0   0   0, ARTICLE CONTAINS CONSUMER GOOD(S)    FALSE   FALSE   0   0,  FALSE   FALSE   FALSE   FALSE   FALSE, FALSE, , , , , , , , , , , , , , , , 
C   QTM0039065  7021130 X6  100013177,  Example, , Example  VIC 3136    AU  61397233661 FALSE,  TRUE    FALSE, N    0, FALSE, FALSE, FALSE  Example Example, , Brisbane QLD 4000    AU  Example Example Example, , Example  QLD 4211    AU, Example TRACKADV

I only need data from rows C. Is there a quick way to either delete all rows starting with 'M' and 'A', or, to ignore these rows in a script?

If I remove all of the M & A rows manually from the target file I can use this to get the data I want, but because the row lengths are different I'm having trouble using this method regardless of new row character.

if (($handle = fopen("test.csv", "r")) !== FALSE) {
    while (($data = fgetcsv($handle, 1000, ",", "\n")) !== FALSE) 
    {
        echo $data[0] . " - ". $data[1] . " - ". $data[4] . "<br/><hr>" ;
    }

    fclose($handle);
}

Answer 1

One way would be to

preg_replace_all('#^([MCA])\s#im',"$1,",$file);

and then parse it as a normal CSV

while($line=fgetcsv($f))
{
    if($line[0]=="M" || $line[0]=="A") continue;
    /* ... */
}

If you're using PHP >=5.3, you can use the other way: str_getcsv

while($line=fgets($f))
{
    if(preg_match('#^[MA]\s#i',$line)) continue;
    $line=str_getcsv(substr($line,2));
    /* ... */
}

One downside of this method is if your CSV contains something like "multiline \\n column" , it will break.

And about the variable-length problem, you may be interested to know that fgetcsv only need one parameter to work (PHP >=5); with all the other parameters left to default, it will read every CSV-line in full length.