Using PHP to take the first line of a CSV file and create a MySQL Table with the data

Question

I am trying to take a rather large CSV file and insert it into a MySQL database for referencing in a project. I would like to use the first line of the file to create the ta

Answer 1

Try casting the value and comparing it with the original one:

define('DECIMAL_SEPARATOR', '.');

switch ($fieldList[$i])
{
    case (string)(int)$fieldList[$i]:
        $typeInfo = (strpos($fieldList[$i], DECIMAL_SEPARATOR) === false) ? 'int(11)' : 'float';
        break;
    case (string)(float)$fieldList[$i]:
        $typeInfo = 'float';
        break;
    default:
        $typeInfo = 'varchar(80)';
        break;
}

Additionaly, check for the presence of decimal separator in the first case for numbers that are round, yet they have the decimal fraction part.

Answer 2

Building on Nouveau's code you could do this

for($i = 0; $i <= count($fieldList); $i++)
{
    if (is_numeric($fieldList[$i]))
    {
        if (strpos($fieldList[$i],'.') !== false){
            $fieldList[$i] = (int)$fieldList[$i];
        }else{
            $fieldList[$i] = (float)$fieldList[$i];
        }
    }

    switch(gettype($fieldList[$i])) {
        case 'integer':
            $typeInfo = 'int(11)';
            break;
        case 'float':
        case 'double':
            $typeInfo = 'float';
            break;

        case 'string':
            $typeInfo = 'varchar(80)';
            break;
        default:
            $typeInfo = 'varchar(80)';
            break;
    }
if(gettype($fieldList[$i]) != NULL) echo "\t".'`'.$i.'` '.$typeInfo.' NOT NULL, --'.gettype($fieldList[$i]).' '.$fieldList[$i]."<br/>\n";

}

That works, note the addition of "case 'double':" in the switch, but there may be a better way to do the int/float check as they would only work with standard uk/us numbers.

Answer 3

<?php

// GENERATE TABLE FROM FIRST LINE OF CSV FILE

$inputFile = 'file.csv';
$tableName = 'file_csv';

$fh = fopen($inputFile, 'r');
    $contents = fread($fh, 5120); // 5KB
fclose($fh);

$fileLines = explode("\n", $contents);

$fieldList = explode(',', $fileLines[0]);
echo 'CREATE TABLE IF NOT EXISTS `'.$tableName.'` ('."<br/>\n";
for($i = 0; $i <= count($fieldList); $i++)
{

    if(strlen($fieldList[$i]) == 0) $typeInfo = 'varchar(80)';
    if(preg_match('/[0-9]/', $fieldList[$i])) $typeInfo = 'int(11)';
    if(preg_match('/[\.]/', $fieldList[$i])) $typeInfo = 'float';
    if(preg_match('/[a-z\\\']/i', $fieldList[$i])) $typeInfo = 'varchar(80)';

    echo "\t".'`'.$i.'` '.$typeInfo.' NOT NULL, -- '.gettype($fieldList[$i]).' '.$fieldList[$i]."<br/>\n";
}
echo '  PRIMARY KEY  (`0`)'."<br/>\n";
echo ') ENGINE=MyISAM  DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci AUTO_INCREMENT=1 ;';

Answer 4

Using regular expressions will give you a flexible solution. If you need to detect date fields then this is the way to go.