- bug-fix applied to expressions using semester and quarter calculation.
- enhanced file operations testing;
- bug fix applied for semester calculation;
- extra-verification added on file statistics determination.
- setting default language to English US
- implemented localization and support for French, Spanish and German languages;
- implemented a complex converter to allow using a JSON file to chain multiple conversions.
- changed short version of newly introduced input parameter (having long name "include-checksum-in-files-statistics") from "t", which was creating a collision for "publish_data_source.py", to a "u" that is safe for all package included scripts.
- made checksum calculation optional for relevant files (both input and output) by adding an input parameter to opt in, which contain warning "may not work with all files content and OS encoding". By circumventing this logic speed is slightly improved with default or specific "No" value for this newly added parameter;
- added a encoding sequence to ensure possible unicode characters in file are not causing fatal error duing checksum calculation sequence;
- switch from a regular print to a timestamped one for recently added feedback targeting internal validations.
- added few additional feedback on internal validation about to happen, to enhance user understanding in case of such values are deemed to be invalid.
- BUG FIX: a default value for log input parameter is properly treated on log folder check (aka check is circumvented);
- added Spanish localisation;
- implemented localization maintenance besides existing compilation + added localization to localization common class;
- reflecting JSON file format support in documentation;
- added an extra verification for the scenario when at least one of the source files is newer than destination "overwrite" decision will be stronger than given "create" (safer to data accuracy).
- extended functionality and now input file and/or output file are aware of Calculated Date expressions, making possible this way to use dynamic name interpreted at run time;
- added "create" as option for "policy-to-handle-hyper-file" input parameter that allows skipping recreating "hyper" file if already exists;
- improved feedback.
- bug fixes on internal code structure.
- extended the sql input parameter to be used in conjunction with Read action, this way a user customized query can be used to extract partial data (besides existing full extract one)
- few internal code and localization messages enhancements and corrections.
- added a data remove feature from Tableau Extract file (Hyper format) by extending one of the input parameters and add an additional one to have the ability to specify a custom SQL to remove data based on customer needs.
- added additional input parameter to have be able to create (or recreate if hyper file already exists) new hyper file and add data from input files OR just use existing hyper file and append data from input files (can be very useful in case of large data sets).
- added support for 2 additional file format in a combination with output file format = Hyper;
- few additional file formats combination checks to avoid runtime errors;
- improved localized feedback captured in log files.
- supported files are now used internally for validation purposes which makes code more maintainable if and when new file types will be implemented with changes in fewer places.
- added support for Parquet file format (full coverage: both input and output).
- extended support for Pickle and JSON file format which now can be used as input files in combination with Tableau Extract (Hyper) as output file;
- field type timestamp properly recognized;
- added a matrix of file combination in documentation (should make usage much clearer).
- tweaked the localization logic that manages language to use based on operating system default region language.
- forced a value to be string converted when logged.
- region language detection improved.
- internal project restructure.
- added localisation (currently covering en_US, it_IT, ro_RO locales);
- operating system current region locale is auto-detected and if
- added support to convert Excel, Pickle and JSON formats into CSV;
- revised internal code towards a more manageable one.
- added input and output (where relevant) file statistics to log;
- code cleanup;
- implemented input validation for most important values;
- corrected internal package name in the setup file;
- updating README file to reflect recent file name change;
- minor code tweaks;
- skipped (13...)
- minor code tweaks
- renamed
publisher.pytopublish_data_source.py;
- renamed
main.pytoconverter.py(apologies for any inconvenience this may create to any user); - various internal code quality improvements;
- internal code change
- just corrected setup.py
- just corrected setup.py
- added a file existing check to circumvent a package crash in case of file not existing;
- added support for large input file;
- added support for upper case months and weekdays besides existing proper case;
- added timing for certain code performance masuring (that is exposing such information to logger);
- bumped version of TableauHypeAPI (former 0.0.9746 => new 0.0.10309 which was released on 2020-03-25);
- added additional script to facilitate publishing a Tableau Extract file (Hyper format) to a specified Tableau Server.
- bug fix to get only non-null unique values analyzed in case of a float column that may hold disguised integers
- internal code tweaks to reduce complexity for one long method
- updated download source URL for Tableau Hyper API
- added boolean data type
- added an optional log file to ease debugging activity when necessary
- bumped version of TableauHypeAPI to 0.0.9746 (which is now available on pypi.org as tableauhyperapi)
- internal code optimizations
- switched default behavior to store 0 values instead of nulls in case of integer determined fields
- just reflecting new version of Tableau Hyper API (former 0.0.8952 => new 0.0.9273)
- additional data type formats recognized:
- date-DMY-dash;
- date-DMY-dot;
- date-DMY-slash;
- date-MDY-medium;
- date-MDY-long;
- datetime-MDY;
- datetime-MDY-micro-sec;
- datetime-MDY-medium;
- datetime-MDY-medium-micro-sec;
- datetime-MDY-long;
- datetime-MDY-long-micro-sec;
- various code optimizations:
- eliminating data type as method parameters and using original class variable;
- moved data type determination to main program logic and out of the Hyper creation one = segregation of duties;
- split analyzing method into 2 separate parts, one for very first unique value and another methods for rest;
- switch to a more standardized class name in the main logic.
- switch the handler of user input parameters from getopt to argparse package (greater flexibility, shorter code, easier future maintenance)
- switch from classic loops to Pandas way of handling data manipulation