csv sample
timestamp,name,amount
1571967208,Rahul,15.7
1571967200,Raju,25.7
code sample
mongoimport --db=crypto --collection=t --type=csv \
--columnsHaveTypes \
--fields="timestamp.date(), name.string(), amount.double()" \
--file="text.csv"
result
Failed: type coercion failure in document #1 for column 'timestamp', could not parse token '1571967208' to type date
You could pre-process your CSV into MongoDB Extended JSON (v2) using jq
with something like:
jq --null-input --raw-input 'input | split(",") as $fields | inputs | split(",") as $values |{($fields[0]): {"$date": {"$numberLong": ($values[0]+"000")}}, ($fields[1]): $values[1], ($fields[2]): ($values[2] | tonumber)}' yourFile.csv > yourFile.json
This would transform your example CSV into:
{
"timestamp": {
"$date": {
"$numberLong": "1571967208000"
}
},
"name": "Rahul",
"amount": 15.7
}
{
"timestamp": {
"$date": {
"$numberLong": "1571967200000"
}
},
"name": "Raju",
"amount": 25.7
}
See jqplay.org example.
and then:
mongoimport --db=crypto --collection=t --file=yourFile.json