Null Values by migration from MySQL to mongo - mysql

Question

I need to migrate some tables from MySQL to mongoDB. After searching the web, for me it looks like an MySQL export to CSV and an import from that CSV to mongoDB should be the fastest and easiest way.
I'm export MySQL using that query:
select * into outfile '/tmp/feed.csv'
FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'
LINES TERMINATED BY ''
from feeds;
But there is one problem.
If an MySQL field is NULL, so the MySQL export writes an \N (or \\N) into the CSV file.
By importing that file, mongoDB imports the \\N as string instead of an NULL value.
The mongoDB import option --ignoreBlanks will not work, becaouse \\N is not "blank" in mongoDB's point of view.
So my question:
1.) how could I avoid exporting NULLas \\N?
or
2.) how could mongodbimport read/interpret \\N as NULL or empty value?
By the way: it's not an option to postprocess the CSV to search and replace the \\N
On possible answer for 1.) could be the modification of the select statement: SELECT IFNULL( field1, "" ) But in this case I have to define and check every column. An export script would not so flexible, if all columns are defined in the select statement.
//Edit: while playing around with that import<->export I found an other problem: date fields, which also interpreted as strings from mongoimport

Answer 1

I would comment rather than adding an answer, but my reputation is still quite low...
What I've done in a project I'm working on is do the migration using a Python script. I have the exported table in a CSV. The code I use looks like this:
import csv
import zip
import pymongo
f = open( filename )
reader = csv.reader( f )
destinationItems = []
The following reads the column names (first row in CSV)
columns = next( reader )
The columns can be put in a tuple that here I call 'keys'. The code is here oblivious of the column names. Each row is then converted to a dictionary ready to be amended to remove (or do something else with -) NULLs.
keys = tuple( columns )
for property in reader:
entry = dict( zip( keys, property ) )
and the following deals with NULL; in this case I remove the entry altogether if found to be 'NULL' in the exported CSV.
entry = { k:v for k,v in entry.iteritems() if ( k in keys and ( v != 'NULL' ) or k not in keys ) }
destinationItems.append( entry )
Update the mongodb instance
mongoClient = pymongo.MongoClient()
mongoClient['mydb'].mycollection.insert( destinationItems )