[teiid-issues] [JBoss JIRA] (TEIID-4520) Support for Exasol Connector

[ https://issues.jboss.org/browse/TEIID-4520?page=com.atlassian.jira.plugin... ] Rafal Korytkowski edited comment on TEIID-4520 at 5/16/18 10:17 AM: -------------------------------------------------------------------- [~rareddy], do you basically disagree with [~shawkins]? Let's try to get it straight. Here are the options for handling NullOrder.LAST as defaultNullOrder: 1." _SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum_" will be rewritten to "_SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum NULLS FIRST_" to match Teiid's expected ordering and returning *NULLs as first*. 2. "_SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum_" will not be rewritten and *NULLs will be returned last*. a) "_SELECT IntNum FROM BQT1.SmallA ORDER BY convert(IntNum, bigdecimal)_" will return *NULLs as first* since Teiid's internal sort will be applied. b) "_SELECT IntNum FROM BQT1.SmallA ORDER BY convert(IntNum, bigdecimal)_" will return *NULLs as last* since Teiid's will respect defaultNullOrder when doing internal sort. 2a is the current behavior, which I don't think is quite right. IMO 1. would be the most expected behavior, however 2b is also acceptable. was (Author: rkorytkowski): [~rareddy], do you basically disagree with [~shawkins]? Let's try to get it straight. Here are the options for handling NullOrder.LAST as defaultNullOrder: 1." _SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum_" will be rewritten to "_SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum NULLS FIRST_" to match Teiid's expected ordering and returning *NULLs as first*. 2. "_SELECT IntNum FROM BQT1.SmallA ORDER BY IntNum_" will not be rewritten and *NULLs will be returned last*. a) "_SELECT IntNum FROM BQT1.SmallA ORDER BY convert(IntNum, bigdecimal)_" will return *NULLs as first* since Teiid's internal sort will be applied. b) "_SELECT IntNum FROM BQT1.SmallA ORDER BY convert(IntNum, bigdecimal)_" will return *NULLs as last* since Teiid's will respect defaultNullOrder when doing internal sort. 2a is the current behavior, which I don't think is quite right. IMO 1. would be the most expected behavior, however 2b is also acceptable. -- This message was sent by Atlassian JIRA (v7.5.0#75005)