1. 12 5月, 2021 2 次提交
  2. 10 5月, 2021 2 次提交
  3. 17 3月, 2021 1 次提交
  4. 04 3月, 2021 1 次提交
  5. 26 1月, 2021 1 次提交
  6. 20 1月, 2021 1 次提交
  7. 06 1月, 2021 1 次提交
  8. 05 1月, 2021 1 次提交
  9. 31 12月, 2020 1 次提交
  10. 30 11月, 2020 1 次提交
  11. 26 11月, 2020 1 次提交
  12. 11 11月, 2020 1 次提交
  13. 30 10月, 2020 1 次提交
    • V
      chore: preparation for processing of Kafka jdbc connector queries (#679) · 74eb8dfd
      Vlad Ilyushchenko 提交于
      - refactored function cache to parse function signature only once and support array types
      - refactored function parser to use function descriptor instead of parsing string again
      - refactored function parser to match array types correctly
      - refactored function parser not to create constants prematurely but create constant once against to level function. This leads to fewer allocations
      - added generic StrArrayFunction and introduced [] operator for dereferencing string arrays
      - created "current_schemas" function that is required for PG metadata
      - fixed memory leak in PGJobContext
      - created and tested "replace(string, from, to)" function for string arguments. It is required by PG metadata queries. Added tests
      - implemented "function call" dereference, e.g. f().x in ExpressionParser
      - added ability to execute SQL in select clause (doesn't make sense? read details)
      
      PostgresSQL metadata SQLs sometimes look like:
      
      ```sql
      select a, (select a,n,c from x).n, c from y
      ```
      
      We now support this sort of query and resolve '.n' to column name in the cursor.
      
      - fixed tests broken by new JDK release
      74eb8dfd
  14. 29 10月, 2020 1 次提交
  15. 20 10月, 2020 1 次提交
  16. 02 9月, 2020 1 次提交