Redshift is a batch-oriented environment optimized to perform large data loads utilizing the
COPY command. That is not to say that you cannot get something approaching "real time", as we discussed earlier in the book. However, you do need to keep in mind all of the work that is going into the distribution and replication of the underlying data for each transaction you are committing. In addition to this, keep in mind that Redshift will never modify data in-place. Updates will actually delete the data and insert a new row with the updated data. It is best to perform bulk operations with the
COPY command whenever possible.
COPY command will load data in parallel. Take advantage of this by splitting your files into parts in a directory ...