Commits on Apr 9, 2014
  1. implement RenameColumn() ModifyColumn() for VerticaDataUtils

    add the ShallRefreshViews() and related utility methods into VerticalDatUtils to determine if we should refresh views for a give set of conditions.
    Nan Meng committed Apr 9, 2014
Commits on Mar 24, 2014
  1. 1699344 - bug fix, when running single batch mode, Tesla Vertica slav…

    …e complains about absent PK info when there are tables with no changes
    Nan Meng committed Mar 24, 2014
Commits on Mar 20, 2014
  1. fix NULL value in MSSQL to Vertica data copy

    Nan Meng committed Mar 20, 2014
Commits on Feb 14, 2014
  1. Add Vertica extension.

    Change method signatures for IDataUtils ModifyColumn and AddColumn
    passing DataType instead of string data type to have more data type
    info available.
    
    Fix one unit test.
    
    Add helper functions for TableConf.
    
    Fix RefreshView TableName property logic, adding its unit test
    Nan Meng committed Feb 14, 2014
Commits on Feb 3, 2014
  1. Merge branch 'master' of git.csnzoo.com:wayfair/tesla

    Conflicts:
    	TeslaSQL/DataUtils/MySQLDataUtils.cs
    	TeslaSQL/TeslaSQL.csproj
    Nan Meng committed Feb 3, 2014
  2. start adding Vertica extension for Tesla

    Nan Meng committed Feb 3, 2014
  3. @zerobfd

    Fixing bad merge

    zerobfd committed Feb 3, 2014
  4. @zerobfd

    Merge branch 'working'

    Conflicts:
    	TeslaSQL/Agents/Notifier.cs
    	TeslaSQL/DataCopy/MySQLToMSSQLDataCopy.cs
    	TeslaSQL/DataType.cs
    	TeslaSQL/DataUtils/MSSQLDataUtils.cs
    	TeslaSQL/DataUtils/MySQLDataUtils.cs
    	TeslaSQL/TeslaSQL.csproj
    zerobfd committed Feb 3, 2014
  5. @zerobfd

    Hacked together some failure handling for MySQL

    The way the program flow is right now, the Master agent checks the master
    DB first, then the relay, then starts copying. Because of the way we have
    MySQL set up as a master, there is a point in the program where if it
    fails without completing, some data will never be copied over to the relay
    because it resets the time range to copy. I added a column in the MySQL
    CTIDtoTimestamp table that gets a 0 written to it when a line is inserted,
    then in the cleanup init table (the last method in the Master.Run()
    function) it writes a 1. I know this isn't great design, but until I have
    time to rewrite more of it I don't want to mess with the program flow and
    risk breaking our current working Tesla agents, so I'm leaving this one
    until a different sprint.
    zerobfd committed Sep 23, 2013
  6. @zerobfd

    Switching cast from Int64 to Int32

    Not sure why we were originally casting it to an Int64 in the first place,
    although I seem to remember that MySQL stored a lot of things as Int64
    where MSSQL didn't. Anyways, it was throwing cast errors and now it's not.
    zerobfd committed Sep 20, 2013
  7. @zerobfd

    Adding order by table_name

    There are some cast errors popping up after ordering by ordinal_position,
    so we're also ordering by table name first since the code looks like it
    depends on all of the columns in a table being grouped.
    zerobfd committed Sep 19, 2013
  8. @zerobfd

    Specifying column ordering in information_schema

    Even though the columns normally come back from information_schema in the
    correct order, we ran into a place where apparently they're not getting
    called in the correct order, so we're changing the query that
    MySQLDataUtils uses to get the fields and specifying to order by
    ORDINAL_POSITION.
    zerobfd committed Sep 18, 2013
  9. @zerobfd

    Switching DataType to use longs instead of ints

    In MySQL , there's a field called longtext that has the CHARACTER_MAXIMUM_SIZE
    defined as some huge number that's longer than an Int32 can handle, so we're
    changing the struct and a couple of methods to use Int64s and Nullable<Int64>s
    instead.
    zerobfd committed Sep 16, 2013
  10. @zerobfd @zerobfd
  11. @zerobfd @zerobfd

    switching branches

    zerobfd committed with zerobfd Sep 5, 2013
  12. @zerobfd @zerobfd

    cannot rely on mysql cluster to autoincrement like you think it will,…

    … so we handle incrementing the ctid manually now
    zerobfd committed with zerobfd Aug 26, 2013
  13. @zerobfd @zerobfd

    added function to get stopsyncversion from the relaydb instead of cti…

    …ds and pass that into the mysqldatautils
    zerobfd committed with zerobfd Jul 29, 2013
  14. @zerobfd @zerobfd
  15. @zerobfd @zerobfd

    moving the logic that exits early if no errors found to the beginning…

    … of the notifier agent instead of after the email sends
    zerobfd committed with zerobfd Jul 25, 2013
  16. @zerobfd @zerobfd
  17. @zerobfd @zerobfd
  18. @zerobfd @zerobfd
  19. @zerobfd @zerobfd
  20. @zerobfd @zerobfd

    mysql uses the charactermaximumlength on stuff like text where it is …

    …not explicitly set, so account for that in the mysql->mssql datacopy
    zerobfd committed with zerobfd Jul 17, 2013
  21. @zerobfd @zerobfd
  22. @zerobfd @zerobfd
  23. @zerobfd @zerobfd

    adding cleaning up trigger tables

    zerobfd committed with zerobfd Jul 11, 2013
  24. @zerobfd @zerobfd
  25. @zerobfd @zerobfd
  26. @zerobfd @zerobfd
  27. @zerobfd @zerobfd

    finished the first pass

    zerobfd committed with zerobfd Jun 12, 2013
  28. @zerobfd @zerobfd
  29. @zerobfd @zerobfd

    adds and drops work

    zerobfd committed with zerobfd May 21, 2013
  30. @zerobfd @zerobfd
  31. @zerobfd @zerobfd

    almost done with the ddl stuff

    zerobfd committed with zerobfd May 16, 2013