Invalid digit, Value '\\', Pos 0, Type: Long error in Amazon redshift copy command

Question

I'm trying to unload the data from Amazon redshift table and then copying the data from s3 file to redshift by using following command.

unload ('select * from abx.xuz where src =\'vpc\'') to 's3://load/2017/' credentials 'aws_access_key_id=******;aws_secret_access_key=*****' allowoverwrite;

copy abc.abxy from 's3://sumo-processed/2017/' 
credentials 'aws_access_key_id=XXXX;aws_secret_access_key=YY' NULL AS '\\0' TRUNCATECOLUMNS FILLRECORD  maxerror as 255;

But it is not able load all the records. After checking the stl_load_errors table it is showing "Invalid digit, Value '\', Pos 0, Type: Long" error.

How to resolve this error?. SO I can load all recording without skipping?

Solution

I was having similar error "Invalid digit, Value '.', Pos 2, Type: Integer", and the reason was pandas.io.sql.read_sql converting all "INT NULL" columns into "FLOAT"; so I had 10.0s instead of 10s in the dataframe. Then I dump the dataframe to S3, and then I import it from S3 to REDSHIFT using COPY , getting that error because COPY cannot import 10.0 to REDSHIFT's INT4.

Fix: replaced all "INT NULL" fields in the destination REDSHIFT table with "DECIMAL(10,0)"; after that COPY works like a charm.