Deal with timestamp

[maykiwo]

Thanks for your work.
It seems have a problem with timestamp or i've not used correctly your jar

for testing i have got this file
hello|1|2003-03-05 18:09:28.0
world|2|2006-03-05 17:09:28.0

Use the jar like this :
java -jar csv-to-orc.jar -i hello.csv -o hello.orc -s 'structbonjour:string,num:int,create_date:timestamp' -sep '|' -skipcount 0

Dump the orc file :
$ hive --orcfiledump -d -t /user/hdp/test/hello.orc
2017-12-07 09:49:58,366 INFO [main] orc.RecordReaderImpl: Reader schema not provided -- using file schema structbonjour:string,num:int,create_date:timestamp
{"bonjour":"hello","num":1,"create_date":"1970-01-01 01:00:00"}
{"bonjour":"world","num":2,"create_date":"1970-01-01 01:00:00"}

the timstamp is set to linux time. it is the behavior of your jar ?
I have excepted to retreive the timestamp of csv file.

regards

[egor-ryashin]

Added #2

cat 1.csv
time,tag,number
2018-07-09T21:00:00Z,"a",2
2018-07-09T21:00:01Z,"b",3

java -jar target/csv-to-orc-1.0-SNAPSHOT-jar-with-dependencies.jar -i 1.csv -o 1.orc -s 'struct<time:timestamp,tag:string,number:int>' -sep ',' -skipcount 1

bin/hive --orcfiledump -d -t 1.orc
Processing data file 1.orc [length: 411]
{"time":"2018-07-10 00:00:00.0","tag":"a","number":2}
{"time":"2018-07-10 00:00:01.0","tag":"b","number":3}