inkimar / commons_cli_ex Goto Github PK

it should be sufficient to only run 'make' to be able to test the project with the test-database.
target 'all' should do the following.

1. download the example database from Internet Archive ( IA) if it does not exist
2. clean out src-code and output-files
3. build the executable jar
   4 runt the project and generate csv-files

The mdbtools output results to stdout.

It would be nice if the tool would print space separated table names (unless they need quoting?) to stdout when called with the option --list

It would be nice if the tool would output the schema DDL (sqlite / ANSI sql) to std when called with option --schema.

Currently there are some "java.io.IOException: Permission denied" messages which seems related to accessing the filesystem which shouldn't be needed for those options ie --schema and --list.

An Alpine based image with the acess2csv tool available on Docker hub.

Perhaps this distribution could be based on openjdk:8-jre-alpine from Docker Hub - it would be nice, portable and a simple install.

It could support usage when mobilizing data from old Access databases to other formats via CSV exports in the following way:

# download the tool
docker pull inkimar/access2csv

# running it without options will display available options and inform about usage
docker run --rm inkimar/access2csv 

# example usage running a conversion of an access db into CSV files which are stored in the "output" destination directory
docker run --rm inkimar/access2csv -f my_access_db.mdb -d output

The flags should not been displayed ?

https://commons.apache.org/proper/commons-cli/
javadock ->
http://commons.apache.org/proper/commons-cli/javadocs/api-release/index.html

When running the tool with the "-d" option, generated files should be similar to files generated from the "mdb-export" tool:

• files start with a first row with the column field names (if column names contain a comma or a tab ie the field separator character it would need to be quoted or even possibly escaped?)

• a row in a CSV file output should encode missing values using empty values "" instead of the string "null"

• a row in a CSV file should by default quote text fields using " as the quotation character

build releases.
Those releases can be used instead of building the whole project.

now one has to update the SCHEMA-variable to change db.

the same as the input-directory.
the -d should not be mandatory

When the tool is called with either --help or -h it should print its manual, possibly also if incorrectly called.

However, if called with options that make sense, the manual should not be printed to stdout.