Neo4j Etl Versions Save

Fixes deployment so that neo4j-desktop works again.

Fixes #274: When switching to another desktop project the rdbms connection fails Fixes #249: state change of database is not correctly detected

This release contains updates for Neo4j 4.0

• Signed for Neo4j Desktop 1.2.5 with new certificate
• support of Neo4j 4.0 databases incl multi-database
• replaced neo4j-import with neo4j-admin import for bulk import
• added UNWIND option for cypher batching
• added multi-db support for online import
• updated neo4j driver to 4.0.1
• added support for streaming data directly from relational database without intermediate CSV files
• fix camel case property names in UI

Thanks a lot to @conker84 and @mroiter-larus for all your hard work!

Our good friends at The SilverLogic helped us to upgrade the code to React and design a new style and flow.

Bug Fixes

A number of issues were fixed.

• support for new auth approach in MySQL 8.0.4+, upgraded packaged MySQL driver
• fix for Azure SQL database
• fixed CSV quoting
• composite key handling for MSSQL

Something we learned is that extracting metadata from Postgres is quite slow, which is an issue of the Postgres JDBC driver. This issue becomes exacerbated with remote database, so please be aware.

See the article for more details

This release contains mostly bugfixes:

• fix quoting for neo4j-import
• fix bolt auth if no username/password provided
• Windows environment fixes (#116)
• check if Neo4j instance is running before import fix #22 (#67)
• including only MySQL and PostgreSQL drivers in the uber jar fix #46 (#68)
• Replaced ResultSet.getString with ResultSet.getObject to properly extract data from RDBMS.
• stemmer removed, mysql driver to 5.1.38, better datatype conversion (#74)
• better info when setting up a new connection to a RDBMS
• neo4j-admin import with --f parameters fixes #112
• neo4j-shell fails on Windows fixes #114

We're happy to finally release the 1.2.0 version of Neo4j-ETL.

From 1.1.x it added quite a number of features.

Features

• integrated as Graph-App in Neo4j Desktop (ask your neo4j-contact for application-key or [email protected])
• bundled MySQL and Postgres drivers
• allow custom JDBC drivers for all other databases (with Neo4j Enterprise) e.g. MS-SQL, DB2, Oracle, ...
• proper schema support
• include/exclude tables, schemas, columns directly and by patterns

Bugfixes

• handle schema with dashes
• numeric field no longer treated as BIT in Oracle
• handle casing of schema
• fix CSV quoting
• one directory per run
• save mapping files from desktop properly
• better logging
• load csv paths for windows
• stream from command-line invocations
• handle graph databases status

After GraphConnect NYC we did a lot of work to fix various bugs and to enhance the ETL tool adding support for all relational databases with a JDBC driver. We did also some backend optimizations and few changes to the UI.

• Multi schema support, the ETL tool now can “parse” more than one schema at time
• Additional driver support. The ETL tool comes with two embedded JDBC drivers (for MySQL and PostgreSQL) but you can set up an external JDBC driver for: Microsoft SQL server, IBM DB2 and Oracle (the list is not limited to these drivers but they are the default ones in the combobox). You can add a jar using “--driver” parameter.
• The resulting mapping file can now be written to a file without output redirection using “--output-mapping-file”.
• Fetch size has been added to a default value of 10,000 records. It will be configurable in future releases
• The ETL tool has undergone additional testing with Microsoft SQL sample databases AdventureWorks and WideWorldImporters and a DB2 sample database in addition to the previous tests
• When importing through cypher all fields are now mapped correctly according to their data type
• First version for parallel generation of the CSV. The concurrency will be configurable in future releases
• Now the ETL tool creates a separate directory for each schema/catalog when writing the CSV files
• Schema names are well-separated from table names. The ETL tools doesn't rely anymore on splitting names with "." in order to separate the schema name from the table name. These changes reflects also on the mapping.json file where the name of the schema is explicitly written.

We're extremely happy to provide you the latest preview release of Neo4j-ETL.

Thanks to our partners at Larus IT the neo4j-etl tool made great progress in the last few months.

We now bundle the drivers for postgres and mysql and added support for any database JDBC drivers (MS SQLServer, Oracle, DB2, ...) for Neo4j Enterprise.

Fixed issues:

• better data type conversion (esp. for oracle)
• handling multiple schemas with duplicate table names
• escape unusal table name symbols
• handle unknown data types
• import into running databases via bolt-connector, cypher-shell and neo4j-shell

Thanks a lot to @inserpio for all the work and @praveenag for the review.

Update to Neo4j 3.1.4

This release adds support for Postgres and Oracle databases.

Please also see the updated documentation

It contains now also step by step instructions for setting up the source-databases with Docker, e.g. for testing.

All you have to do is to download the JDBC driver for the database and add it to the lib folder and use the correct parameters for the command line tool.

• fix neo4j-etl script for unix and windows to include database drivers from lib directory
• support for unsigned / tiny-int-fields thanks @kbsali

We're happy to provide the first release of neo4j-etl-components.

We would love you feedback on this tool.

Please read the docs and the blog post by @praveenag for some more details.

Currently the tools support mysql but other relational databases with JDBC drivers should work too.

The approach works by generating a graph model from a reasonable entity-relationship-model and importing exported CSV files from the relational database.

Grab the ZIP file for the binary distribution, you'll have to have the zip-distribution of Neo4j-3.0.x installed for this to work.