Parsing the output of SELECT INTO OUTFILE

Question

I'm trying to parse the CSV output of SELECT INTO OUTFILE. (The purpose is to bulk-load the data into an external NoSQL database; the bulk loader requires a specific file format so I'm writing a generator)

My main problem is the handling of escaped characters - specifically, the FIELDS TERMINATED BY character. MySQL does not seem to escape the delimiter when it occurs in a column value.

The export options of my query are as follows:

FIELDS
    TERMINATED BY ','
    ENCLOSED BY '"'
    ESCAPED BY '\\'
LINES
    TERMINATED BY '\n'

I get CSV lines like:

"value1","some, value","another value","value3"

This breaks my CSV line parser that simply splits the line using the FIELDS TERMINATED BY character.

ie my parser produces the following column values:

• value1
• some
• value
• another value
• value3

"some" and "value" should have been parsed as a whole string "some, value"

What is the correct way to parse the output of SELECT INTO OUTFILE?

Anything else that I should worry about given the export options above? My code already handles the following cases within a column value:

(in exact order)

• Treat \\N column value as NULL
• Replace \\" with "

Note:

If anyone can provide the pseudocode of MySQL's LOAD DATA INFILE, that would be fantastic. The exact behavior of that statement is what I'm trying to emulate.

Note:

My parser is written in Java, but I can understand PHP much better

Answer 1

I ended up using OpenCSV ; although I had to write a special handler for NULL value which MySQL dumps as "\\N". Fortunately, my schema does not have any nullable string field so I can safely assume that all occurrences of "\\N" are NULLs.

Thanks to @Doge for reminding me not to reinvent the wheel :)