CHANGELOG.md

Change Log

1.6.4

• Cleaned up several warnings due to Elixir versioning deprecations. Thanks to @michaelst and @getong for contributions.

1.6.3

• Fixed bug where multi_extract did not return {:error, msg} for non existent file. Thanks to Peter Sumskas (@brushbox) for contribution.
• Fixed bug where formula was returned if calculated value was empty. Thanks to Ken Ip (@kenips) for contribution.
• Sheet name added to Xlsxir.get_info/1 and updated formatting. Thanks to Hongseok Yoon (@hongseokyoon) for contribution.
• Updated docs.

1.6.2

• Fixed bug where Xlsxir.get_list/1 was not populating empty cells with nil propely.
• Pattern matching on error cells was widened to include additional use cases. Thanks to Peter Sumskas (@brushbox) for contribution.
• Updated tests, docs and various code styles.

1.6.1

• Fixed bug where Xlsxir.get_cell/2 raised instead of returning nil on non-existing cell. Thanks to @ZombieHarvester for contribution.
• Various documentation updates.

1.6.0

• Huge parsing performance improvement thanks to Alex Kovalevych's (@AlexKovalevych) contribution.
• Ability to choose between parsing in-memory or on the file system added as well as the ability to stream via the Xlsxir.stream_list/2. Thanks to Thibaut Decaudain (@Tricote) for contribution.
• Code improved to better handle complex multi-formatted strings. Thanks to Peter Sumskas (@brushbox) for contribution.
• Bug fix to handle additional date format. Thanks to Sudhir Rao (@sudrao) for contribution.
• Fixes for some xlsx variants and repeatable stream issues. Thanks to @rhetzler for contribution.
• Various error message improvements. Thanks to Craig Lyons (@craiglyons) for contribution.

1.5.2 (not published on Hex)

• Fixed bug that occured when a worksheet was empty. Thanks to Alex Kovalevych (@AlexKovalevych) for contribution.
• Changed get_cell/1 to return nil if the requested cell doesn't exist. Thanks to Peter Sumskas (@brushbox) for contribution.

1.5.1

• Removed Timex dependency. Thanks to Paulo Almeida (@pma) for contribution.

1.5.0

• Xlsxir requires Elixir v1.4+ with this update
• Added ability to extract only a given number of rows from a worksheet via Xlsxir.peek/3. Thanks to Ali Tahbaz (@tahbaza) for contribution.
• DateTime type values are now converted to an Elixir Naive DateTime type upon extraction. Regular Date types are still converted to Erlang :calendar.date() type. Thanks to Ali Tahbaz (@tahbaza) for contribution.
• A bug in convert_char_number/1 was fixed to allow support for floats with scientific notation in them. Thanks to Daniel Parnell (@dparnell) for contribution.
• Minor bug fixes and documentation updates.

1.4.1

• Added parsing support for time values. Thanks to Edgar Cabrera (@aleandros) for contribution.
• Fixed bug that prevented worksheet ETS tables from closing. Thanks to Alex Kovalevych (@AlexKovalevych) for contribution.
• Minor documentation updates.

1.4.0

• Xlsxir.extract/3 and Xlsxir.multi_extract/3 now parse all worksheets of the file given by default, returning a list of tuple results (i.e. [{:ok, table_1_id}, {:ok, table_2_id}, ...]). See updated docs for more detail. Thanks to Alex Kovalevych (@AlexKovalevych) for contribution.
• Fixed bug where the string(s) from merged cells that contained multiple formatting leaked into other cells thereby corrupting other rows of data.
• Sorted cell attribute keys to ensure consistent pattern matching. Thanks to Alex Kovalevych (@AlexKovalevych) for contribution.
• Updated documentation to reflect changes and added additional doc tests.

1.3.6

• Added boolean value support. Thanks to Pikender Sharma (@pikender) for contribution.
• Added support for data type inlineStr.
• Xlsxir.extract/3 and Xlsxir.multi_extract/3 now return {:error, reason} instead of throwing an exception when an invalid file type or worksheet index are provided as arguments.
• Changed the way file paths are validated prior to parsing. It no longer matters whether or not the extension is .xlsx. As long as it is a valid file, Xlsxir will attempt to parse it.
• Refactored Unzip.delete_dir/1 for simplification.
• Minor documentation updates.

1.3.5

• Fixed bug where unnecessary cells with nil values were added to worksheets with rows containing data beyond column "Z".

1.3.4

• Fixed bug related to parsing a worksheet containing conditional formatting. Thanks to Justin Nauman (@jrnt30) for contribution.
• Fixed bug where row number was erroneously represented as a string (instead of integer) in the ETS table causing Enum.sort to not work as expected on larger files.
• Minor documentation updates.

1.3.3

• Minor bug fixes.

1.3.2

• Fixed bug where dates in the year 1900 were off by one day due to the fact that Excel erroneously considers the year 1900 a leap year.
• Minor documentation updates.

1.3.1

• Fixed issue where empty cells were skipped. Empty cells will now be represented as nil. For example, if cells "A1" = 1, "B1" = 2, and "D1" = 4, Xlsxir.get_list/1 would return [[1, 2, 4]]. The same situation will now return [[1, 2, nil, 4]] to account for the fact that cell "C1" was empty.
• Minor updates to documentation to reflect change.

1.3.0

• Added ability to parse multiple worksheets via Xlsxir.multi_extract/3 which returns a unique table identifier for each ETS process created, enabling the user to access parsed data from multiple worksheets simultaneously.
• Created an Xlsxir.TableId module which controls an agent process that temporarily holds a table identifier during the extraction process.
• Refactored Xlsxir access functions to work with Xlsxir.multi_extract/3 whereby a table identifier is passed through the various functions to specify which ETS process is to be accessed.
• Refactored Xlsxir.SaxParser, Xlsxir.ParseWorksheet and Xlsxir.Worksheet modules to support new functionality.
• Refactored Xlsxir.ParseWorksheet to ignore empty cells.
• Updated documentation and tests
• Fixed a few minor bugs that were generating warning messages.

1.2.1

• Removed Ex-Doc and Earmark dependencies from Hex.
• Added Change Log link to Hex.
• Minor doc changes and bug fixes.

1.2.0

• Added Xlsxir access function Xlsxir.get_mda/0 that accesses :worksheet ETS table and returns an indexed map which functions like a multi-dimensional array in other languages.

1.1.0

• Modified the way rows are saved to the :worksheet ETS table. Replaced the generic index with the actual row number to allow for performance imporovement of supporting Xlsxir access functions.
• Refactored Xlsxir access functions to improve performance.
• Created Xlsxir.get_info/1 function which returns number of rows, columns and cells.
• Various minor modifications to docs.

1.0.0

Major changes in version 1.0.0 (non-backwards compatible) to improve performance and incorporate new functionality, including:

• Refactored the Xlsxir.Unzip module to extract .xlsx contents to file instead of memory to improve memory usage. The following functions were created to support this functionality:
  • Xlsxir.Unzip.extract_xml_to_file/2 - Extracts necessary files to a ./temp directory for use during the parsing process
  • Xlsxir.Unzip.delete_dir/1 - Deletes './temp' directory and all of its contents
• Implemented Simple API for XML (SAX) parsing functionalty via the Erlsom Erlang library to improve performance and allow support for large .xlsx files. The SweetXml parsing library has been deprecated from Xlsxir and is no longer utilized in v1.0.0.
• Implemented Erlang Term Storage (ETS) for temporary storage of extracted data.
• Replaced option argument from the initial extract function (Xlsxir.extract/3) with timer which is a boolean flag that controls Xlsxir.Timer functionality. Data is no longer returned via Xlsxir.extract/3 and is instead stored in an ETS process.
• Implemented various functions for accessing the extracted data:
  • Xlsxir.get_list/0 - Return entire worksheet data in the form of a list of row lists
  • Xlsxir.get_map/0 - Return entire worksheet data in the form of a map of cell names and values
  • Xlsxir.get_cell/1 - Return value of specified cell
  • Xlsxir.get_row/1 - Return values of specified row
  • Xlsxir.get_col/1 - Return values of specified column
• Implemented Xlsxir.close/0 function to allow the deletion of the ETS process containing extracted worksheet data to free up memory.
• Implemented Xlsxir.Timer module for tracking elapsed time of extraction process.
• Changed cell references from atoms to strings due to Elixir atom limitations (i.e. :A1 to "A1").
• Updated documentation and testing to incorporate changes.

0.0.5

• Minor bug fixes and documentation updates.

0.0.4

• Expanded coverage of Office Open XML standard numFmt (Standard Number Format). The formatCode for a standard numFmt is implied rather than explicitly identified in the XML file.
• Implemented support for Office Open XML custom numFmt (Custom Number Format) utilizing the formatCode explicitly identified in the XML file.
• Added Number Styles documentation covering standard and custom numFmt and how to manually add an unsupported numFmt.
• Fixed issue resulting when no strings exist in a worksheet and therefore there is no sharedStrings.xml file (:file_not_found error).

0.0.3

• Fixed issue related to strings that contain special characters. Refactored Xlsxir.Parse.shared_strings/1 to properly parse strings with special characters.

0.0.2

• Refactored Xlsxir.Parse functions to improve extract performance on larger files.
• Expanded documentation and test coverage.
• Completed Xlsxir.ConvertDate module.

0.0.1

• Initial draft. Functionality limited to very small Excel worksheets.
• Xlsxir.ConvertDate functionality incomplete.