Kexi/Migration

Data Migration And Sharing.

Note

TODO: Move content from [1]

• General
• Junior Jobs

File formats

• CSV
• MDB
• ODB
• Spreadsheet
• SQLite
• Fixed Width Text

Servers

• MySQL
• PostgreSQL

Details

There are two kinds of data migration:

• Migration for entire projects.
• Data migration related to a single data table (importing, exporting).

Project Migration

Project migration allows to migrate the database schema and table data in one go. It's currently possible to migrate a MySQL, PostgreSQL or MS Access database to a SQLite database, using either the keximigratetest program or using File->Import Database menu command.

See also: Migration API

Data Migration

Text tabular data can be imported from file (Data->Import Data From File), exported to file (Data->Export Data to File) or pasted from the clipboard (Edit->Paste Special) or copied to the clipboard (Edit->Copy Special).

Text formats for tabular data

There are a number of text formats for tabular data. See MS Excel Import formats for naming scheme.

• Comma Separated Values (CSV) Data
• Fixed Width Text

See also http://en.wikipedia.org/wiki/Flat_file_database

Text formats for tabular data

• Comma Separated Values (CSV) Data
• Fixed Width Text

Data Sharing

By Data Sharing we usually mean dynamic data exchanging features that do not require to importing and exporting. For example: connecting with external (not KDb-derived) database to share some data between KEXI project and that external data source.

Cross-DB-Engine Queries

 jstaniek> mart: btw, another hard way is to add simple querying
 engine on top of KDb (not so hard since we've got schemas parsed
 out  of query statements) and retrieve simple one-table records from
 mdbtools and then combine them using our engine. Advantage:
 we can later even provide a way for making relations for loosely
 connected data sources, as eg. , say, google result set and mysql
 dictionary

 mart> jstaniek: I have had some ideas about 'partial evaluation'
  of SQL before partial evaluation - so I could write SELECT int1, int2,
  str1,str2 FROM t WHERE int1<int2 and string_edit_distance(str1,
  str2) < 5, so the DB goes away and finds records where int1<int2
  then KEXI would execute some plugin-loaded function to filter
  the rest etc...

 jstaniek> btw, it is great tool for data integration... only not
  so robust...  but who cares... that's ALMOST ready!

 mart> jstaniek: it is?

 jstaniek> mart: we've got a lists of builtin (provided at the backend)
  functions, so you iterate through
  PARSE tree of the query and ... split the query to "backed" and
  "frontend" parts...
  (yes, we know these things could be nicer on server level, not client...)
  mart, jan 2005

Speaking of which, it might not be a bad idea to store these functions in Java JAR files. JNI or CNI calls could be made to these functions using GCC's Java support (if available). The JAR could also be deployed on a PosgreSQL server with pljava, allowing the functions to be executed on the server for greater performance.

By the way: a word from MSA team

From Access 12's new data engine article:

[..] Jet is unique in Microsoft because it provides both a data storage mechanism with a heterogeneous query processor. This ability to run queries across a number of different data stores is a key reason people use Access, so it was super important for us to retain that ability going forward. The SQL team is de-investing in Jet, and telling developers to code against SQL Server Express, but as it lacks the heterogeneous query functionality we needed, Access stuck with Jet. [..]

Links

• VistaDB: Migration Wizard Screnshots - usable to learn what options are used when migrating real-world projects.