I'm trying to feed a mongoDB(version 3.2.4) with arcs from google ngrams but it's taking too much time even with a decent configuration (MBP Mid 2014, 2.2 GHz Intel Core i7, 16gb).
For each one of the 8,9million rows in the original file, I create a doc and bulk.insert(doc); it into an unordered bulk.
After inserting 500 of them, I do bulk.execute() and repeat this process until all of them are added to the DB, which never happens, since it takes more than a minute per thousand doc's to be inserted and I sure don't have 8,9k minutes to wait for it.
here goes the code:
MongoClient.connect(url, function (err, db) {
if (err) {
console.log('Unable to connect to the mongoDB server. Error:', err);
} else {
//connected
console.log('Connection established to', url);
var bulk = db.collection('bigrams').initializeUnorderedBulkOp();
const rl = readline.createInterface({
input: fs.createReadStream(path+filename)
});
rl.on('line', function (line) {
var stringArray = line.split("\t");
var firstPart = stringArray[0]+'\t'+stringArray[1]+'\t'+stringArray[2]+"\t";
var head_token = stringArray[0];
var syntatic_ngram = stringArray[1].split(" ");
var total_count = stringArray[2];
var counts_by_year = line.replace(firstPart,'').split("\t");
var doc = {
"head token" : head_token,
"syntatic ngram" : syntatic_ngram,
"total count" : total_count,
"counts by year" : counts_by_year
};
count++;
bulkCount++;
if (bulkCount == bulkSize) {
console.log("BulkSize reached. Executing...");
bulk.execute();
bulkCount = 0;
}
console.log("bulk inserted count:"+count);
bulk.insert(doc);
});
rl.on('end', function(){
bulk.execute();
db.close();
});
}
});
P.S.:The first 10k insertions, i.e., first 20 bulk.execute() are processes really fast <1m. After that processing time goes increasing. ):
Any sugestions? Thank you.