regarding <https://github.com/datalineio/dataline/...
# contributing-to-airbytem
Michel
09/14/2020, 10:36 PM
u
user
09/14/2020, 10:36 PMwe should check what behavior a postgres tap has
u
user
09/14/2020, 10:37 PMif singer doesn't talk about an encoding scheme for '\n' then we need something like that.
u
user
09/14/2020, 10:37 PM@Jared Rhizor (Airbyte)
u
user
09/14/2020, 10:39 PMdo any taps fail to encode newlines to \n that we know of?
u
user
09/14/2020, 10:39 PMwe need to check if the one we implemented are doing something special
u
user
09/14/2020, 10:40 PMThey can not have a requirement like that without explaining what you should do if there is a \n in the data you;re writing
u
user
09/14/2020, 11:39 PMis anyone looking a this?
u
user
09/15/2020, 12:10 AMsorry didn't see this
u
user
09/15/2020, 12:10 AMI changed the sql init file for the acceptance tests to have a record with a newline:
u
user
09/15/2020, 12:10 AM Copy code
(
3,
'jared
newline'
),u
user
09/15/2020, 12:11 AMtests passed and it replicated the same record that was inserted into the destination postgres
u
user
09/15/2020, 12:11 AM Copy code
| id|name |
+----+---------------------+
| 3|jared{lf} n...|
+----+---------------------+u
user
09/15/2020, 12:13 AMmore legibly:
u
user
09/15/2020, 12:13 AMI added
System.out.println("sourceTableRecord = " + sourceTableRecord.get("name"));u
user
09/15/2020, 12:13 AMand it outputs:
u
user
09/15/2020, 12:14 AM Copy code
sourceTableRecord = sherif
sourceTableRecord = charles
sourceTableRecord = jared
newline
sourceTableRecord = michel
sourceTableRecord = johnu
user
09/15/2020, 12:14 AMso it looks like postgres at least is handling escaping correctly for new lines
u
user
09/15/2020, 12:15 AMok
u
user
09/15/2020, 12:15 AMso should just disregard that issue?
u
user
09/15/2020, 12:16 AMI think so. If someone breaks this behavior in a tap., it should represent an exceptional breaking case of that tap
u
user
09/15/2020, 12:16 AMok