Hi Enrico, 1. It should but I fear it is never used 2.
Side note the toUpper in https://github.com/apache/openjpa/blob/0e32d4981abca26ea97d8e66c643b2bf5162d1e9/openjpa-jdbc/src/main/java/org/apache/openjpa/jdbc/schema/Table.java#L335 is very weird but we can likely make the map case insensitive for some DB (going through the dict probably). Reserved words can be set in the dictionary thanks the variable of the same name (and you likely want to set system tables too), you can have a quick look to https://github.com/apache/openjpa/blob/08bb3a5a0e7573622aa6dc25763cb6513095093b/openjpa-jdbc/src/main/java/org/apache/openjpa/jdbc/sql/IngresDictionary.java#L144 Romain Manni-Bucau @rmannibucau <https://twitter.com/rmannibucau> | Blog <https://rmannibucau.metawerx.net/> | Old Blog <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book <https://www.packtpub.com/application-development/java-ee-8-high-performance> Le jeu. 16 juil. 2020 à 10:08, Enrico Olivelli <eolive...@gmail.com> a écrit : > Hi, > I am working on contributing a DBDictionary for HerdDB ( > https://github.com/diennea/herddb). > > Currently in HerdDBDirectory I have to set > setDelimitIdentifiers(true); > this is because HerdDB uses Apache Calcite and Calcite has a long list of > reserved words like 'User', 'Value'. > > In some real applications this configuration is good and it works very well > but I am not able to make OpenJPA integration tests pass (using > -Ptest-custom profile), in particular the first show stopper is in > TestPersistence. > > I have found this little problem about SelectImpl, the patch is trivial, I > just have to add a test case. > https://github.com/apache/openjpa/pull/68 > > But now I have a harder problem. > > With setDelimitIdentifiers(true) TestPersistence fails on the second test > method (the first passes, testQuery) because SchemaTool does not understand > that the column AllFieldTypes#arrayOfStrings is already present on the > table. > > This is because in SchemaTool at this line: > > https://github.com/apache/openjpa/blob/master/openjpa-jdbc/src/main/java/org/apache/openjpa/jdbc/schema/SchemaTool.java#L608 > > the variable colName holds this value: `arrayOfStrings` (delimited, as the > DBDictionary as delimitAll=true) but dbTable variable (that represents DB > metadata) contains all column names delimited but lowercased (as HerdDB > returns all metadata in lower case) and dbTable.getColumn returns null, > because when we compare delimited DBIdentifiers the comparison is not case > sensitive. > > I have played a lot with DBDictionary#schemaCase > and DBDictionary#delimitedCase, forcing SCHEMA_CASE_LOWER, but each > combination leads to errors. > > I have tried to tweak the generation of DBIdentifiers in the > various AbstractFieldStrategy implementations but anything I touch I break > some existing test case. > > I have also tried to work with 'reservedWords' but it looks like that the > reservedWordsSet is not taken into account when writing DDL and DML (I > didn't test SELECTs) code (looking at RowImpl#getInsertSQL for instance) > > So here a couple of questions for you experts: > 1) Is it good to set setDelimitIdentifiers(true) ? I don't find any > DBDictionary that sets this feature on > 2) If I can't delimit every identifier, is there a way to tell a list of > reserved words to be always delimited ? shall I simply > tweak DBDictionary#toDBName ? > > > Thanks in advance > Enrico >
