Write speed difference - arangodb

Question

I'm doing some testing on arangoDB, and when using the web UI, I want to test the insert speed of 500000 docs
I used this query in the web UI, code is here
FOR x IN 1..500
FOR i IN 1..1000
INSERT {'name': 'hello' ,'age': i } INTO users
This takes roughly 9 seconds on my machine. However, when I use the Python Arango Driver, it takes about 1.9 seconds per 1000 inserts.
Is anyone familiar enough with the python driver to elaborate on why this is?

Answer 1

You are doing different things. Your query in the UI fires more or less just one http request to the server. When you run the insert in a loop, then you do a http request for every item. If you prefer not to use AQL, then you can still send multiple documents in one request. To get an idea look at the following code:
( () => {
const internal = require("internal")
const db = internal.db
const time = internal.time
const print = internal.print
let start_time = undefined
let col = undefined
const col_name = "users"
// one http request
print("aql insert")
db._drop(col_name)
col = db._create(col_name)
start_time = time()
db._query(`
FOR x IN 1..50
FOR i IN 1..1000 INSERT
{'name': 'hello' ,'age': i }
INTO ${col_name}
`)
print(time() - start_time)
// one request for every document
print("rest insert")
db._drop(col_name)
col = db._create(col_name)
start_time = time()
for(var x = 1; x <= 50; x++) {
for (var i = 1; i <= 1000; i++) {
col.insert({ name: 'hello', age: i})
}
}
print(time() - start_time)
// 50 http requests
print("rest batched insert")
db._drop(col_name)
col = db._create(col_name)
start_time = time()
let batch = []
for (var i = 1; i <= 1000; i++) {
batch.push({ name: 'hello', age: i})
}
for (var x = 1; x <= 50; x++) {
col.insert(batch)
}
print(time() - start_time)
})()
The code can be executed in arangosh with require("internal").load(<path to file>).
possible output:
127.0.0.1:8529#_system> require("internal").load("/home/ulf/insert-example.js")
aql insert
3.2375659942626953
rest insert
13.451776504516602
rest batched insert
3.7614316940307617