34490225/read-last-n-lines-of-csv-file-in-php

<?php // http://stackoverflow.com/questions/34490225/read-last-n-lines-of-csv-file-in-php

// include __DIR__ .'/__bootstrap__.php';

define('SOURCE_CSV_FILE', 'P:/temp' .'/test.csv');

$reqdlastLines      =  125;  // how many lines line we want

$estLineLen         = 1024;  // guess at the line length

$avgLineCnt         =   20;  // number of lines to read to estimate the length
$avgLineCntStartPos =   10;  // number of lines to skip at the start of the file

$avgLineLen         =   -1;  // measured line length from file records

/* --------------------------------------------------------------------------
 * Output array...
 */
$lastRecords = array();

/* --------------------------------------------------------------------------
 * The source file:
 *   o Information
 *   o the actual file handle
 */
$finfo = new SplFileInfo(SOURCE_CSV_FILE);
$file = fopen(SOURCE_CSV_FILE, 'rb');

/* ---------------------------------------------------------------------------
 *  Get an estimate of the average line length
 */
// skip first few records...
for ($recNo = $avgLineCntStartPos; $recNo > 0; $recNo--) {
    $ignore = fgets($file);
}
// estimate the record length based on about 20 records...
$totLen = 0;
$recCnt = 0;
for ($recNo = $avgLineCnt; $recNo > 0; $recNo--, $recCnt ++) {
    $rec = fgets($file);
    $totLen += strlen($rec);
}

// get the average line length
$avgLineLen = max($estLineLen, intval($totLen / $recCnt));

/* ---------------------------------------------------------------------------
 * Position the file pointer to the start of a record near the end of the file
 */

// how far from the end of the file we should be...
$seekPos = $avgLineLen * $reqdlastLines;
$allOk = fseek($file,  $seekPos * -1, SEEK_END); // Must be *negative* position

// now find the start of the next line...
// by finding first none control character...

$readCh = chr(0);
$readChCnt = 0;
while ($readCh !== false && $readCh < chr(32)) {
    $readCh = fgetc($file);
    $readChCnt++;

    if ($readChCnt > $avgLineLen) { // what happened?
        die('looking for a record: read characters: '. $readChCnt);
    }
}

//  and the read rest of this record to get to get to the start of the next record...
$rec = fgets($file);

/* ---------------------------------------------------------------------------
 *  Now we start processing records... yeah!
 *
 *  There needs to $reqdLastLines and no more...
 */
while (!feof($file)) {
    $record = fgetcsv($file);
    if ($record === false) { // end of the file - 'belt and braces'
        break;
    }

    // ensure there are 'required last records' in the array an no more...
    if (count($lastRecords) >= $reqdlastLines) {
        array_shift($lastRecords); // lose earliest record
    }

    $lastRecords[] = $record;
}
// we are done...
fclose($file);

/* ---------------------------------------------------------------------------
 *  Process the last records...
 *
 *  Write to file etc. Here I just dump them
 */
// statistics..
var_dump($avgLineLen, $seekPos, $finfo->getSize(), $lastRecords);