DataExchange Library Release Notes
The first public release of the DataExchange library was version 18.104.22.168, release date July 10, 2020.
Data Exchange Roadmap
The DataExchange library is under active development. The following new features are on the roadmap of the DataExchange library:
Creation of OpenAPI specifications for model-based REST APIs
Automated creation of application databases based on the
Adding these application databases as an additional data source to the (synchronous) DataExchange read and write methods next to JSON, XML, CSV and Excel files.
Exposing these application databases to external applications via dedicated application database-specific API services. This will also allow for asynchronous reading and writing to such application databases from within an AIMMS model.
New Features and Bug Fixes
- 22.214.171.124 [27-03-2023]
Allow JSON documents to expand relative JSON pointers in place
Filter unnecessary parameter schema from generated OpenAPI client code
If possible, provide a more descriptive name for media type schema in generated OpenAPI client code
- 126.96.36.199 [22-03-2023]
Add support in DEX for keeping sessions alive for task REST service in cloud
- 188.8.131.52 [22-03-2023]
64-bits fields in a Parquet file could lead to runtime errors when exceeding
When converting string fields to numeric parameters in the model when reading Excel, partially successful conversions where unconditionally accepted potentially resulting in truncated numerical values. Now partially successful conversions are only accepted when the remainder of the string fields starts with a white space character
Data Exchange runtime errors when reading a file now print a context where the error occurred (e.g. Excel workbook, sheet, row and column)
Fix issue in JSON schema support where an array of arrays would result in a duplicate index in the generated library
- 184.108.40.206 [15-03-2023]
Missing columns in row-based formats that bind to an index are now reported as an error
The error message about mismatching dimensions has been extended with showing the currently bound dimensions
Mapping nodes with duplicate names are now reported as an error
Reading from files with filenames with special characters would fail on Windows
Error messages generated when reading specific row-based formats are now properly propagated and reported
- 220.127.116.11 [09-03-2023]
dex::ReadAllMappingsnow reads all mappings from the
- 18.104.22.168 [07-03-2023]
String fields in an Excel file mapped to a numeric field would be skipped; they are now converted when possible, or produce a runtime error otherwise
- 22.214.171.124 [28-02-2023]
Indices bound via
implicity-binds-toattribute would not always be carried over to parent node to allow usage in sibling nodes
- 126.96.36.199 [20-02-2023]
Trim normalization will now also trim FEFF BOM characters
Labels will be trimmed from FEFF BOM characters before being added to sets
- 188.8.131.52 [12-02-2023]
Dataset mappings generated now also include an Excel mapping that writes sheets regardless of whether or not data is available for that sheet
- 184.108.40.206 [07-02-2023]
Added support for XML request and response bodies in generated OpenAPI clients
- 220.127.116.11 [04-02-2023]
Added trimming leading and trailing spaces off strings as a new string normalization method.
- 18.104.22.168 [30-01-2023]
Empty procedures for JSON schema inadvertently omitted duplicate module prefixes when these occurred in generated identifier names to be emptied
- 22.214.171.124 [09-01-2023]
Data pages for identifiers in the DEX library could cause the extraction of
.aimmspackfiles to fail when the DEX library was included in the
.aimmspack. The publishing process of libraries to the library repository will now automatically remove all data pages.
- 126.96.36.199 [29-12-2022]
no-diacriticsas an additional normalization option next to
- 188.8.131.52 [27-12-2022]
ExcelRootNodefor outputting empty non
This release does no longer support AIMMS versions prior to 4.88
- 184.108.40.206 [05-12-2022]
OpenAPI client code now supports multi-part request bodies
..._itersets generated to add an extra dimension to identifiers for JSON array properties, are now a subset of
Issue a warning for
ColumnNodesin a row-based format mapping (CSV, Excel, Parquet) that cannot be mapped onto a column in a data source during read
dex::ReadFilecall in generated API callbacks to not stop the execution flow when reading faulty responses
- 220.127.116.11 [02-12-2022]
Optional query parameter arguments in generated API calls will only be added as query parameter to the URL if their value is non-default
dex::schema::GenerateClientFromOpenAPISpec()will now generate a library on disk, which can be directly included into your project. Using the generated runtime library directly was often problematic because it is impossible to create parameter with an index domain referring to indices from the runtime library or using sets from the runtime library in the range of element parameters.
Date fields from a Parquet file can now be translated to labels of a regular set, or as values of an element parameter with a regular set range.
- 18.104.22.168 [29-11-2022]
Tab characters in label names were not accepted and would cause a crash, all characters < 32 in label names are now replaced by spaces
- 22.214.171.124 [28-11-2022]
Prevent warning for string parameter passed as handle to external function
AimmsCSVMappingmappings would not accept iterative-reset attribute
Allow only a subset of mappings to be generated with
- 126.96.36.199 [24-11-2022]
Integer-valued headers in Excel files were represented with 5 decimals as a string
Improve double-to-string conversion in the JSON reader to generate the representation using the minimal number of decimals
Add arguments to
dex::client::AddMimePartfor adding headers and encodings to multi-part request bodies
Add support Decimal128, Date32 and Date64 Parquet data types in Parquet reader
Parquet reader would not read Parquet files correctly where not all columns were read into model identifiers
The generated sets
<schemaName>::Instancesare now subsets of the global set
dex::Instancesto make the use of the generated runtime libraries in the main model easier
The generated identifiers
<schemaName>::api::CookieValueare now also dependent on the set
- 188.8.131.52 [15-11-2022]
API keys passed via query parameters did not correctly end up in the URL in api call methods generated by DEX from an OpenAPI specification file
- 184.108.40.206 [11-11-2022]
Reading integer cells from Excel tables into string parameters was not handled correctly
Reading Parquet file containing columns with no data would cause a crash
- 220.127.116.11 [08-11-2022]
Boolean cells from Excel tables were not handled correctly
Improve reading number cells from Excel tables to string parameters, using the minimal number of decimals necessary
Set elements created from integer columns in a Parquet file would cause a crash
- 18.104.22.168 [04-11-2022]
Add methods for computing HMAC and SHA256 digests, base64(-url) encoding and decoding, and url encoding
Add support for binary request and response bodies
EmptyInstancemethods for all generated JSON schema
URL encode the argument values for path parameters in generated
- 22.214.171.124 [28-09-2022]
PATCH curl requests would not send a request body
Better handling of defaults in generated REST API client code to prevent uninitialized data warnings
- 126.96.36.199 [18-09-2022]
Initial release of the REST API client generator from OpenAPI specification files
- 188.8.131.52 [13-08-2022]
ColumnMappingtypes in all row-based mappings, regardless of
name-binds-toattribute. By default, all row-based formats will now leave non-default cells empty.
- 184.108.40.206 [11-08-2022]
Fix string to calendar conversion for CSV and Parquet reading
- 220.127.116.11 [03-08-2022]
Fix string to calendar conversion for Excel reading
- 18.104.22.168 [02-08-2022]
Sheets were read in alphabetical order instead of original order
Write-filter on Excel sheet names was lost during the row-based refactor
- 22.214.171.124 [22-07-2022]
Labels generated from Excel cells with integer values inadvertently contained decimals
- 126.96.36.199 [20-07-2022]
Values from evaluated cells with formulas in Excel files would not be read
- 188.8.131.52 [16-07-2022]
Allow name-binds-to attribute on
- 184.108.40.206 [09-07-2022]
All row-based formats (CSV, Excel, Parquet) refactored to a common code base w.r.t. the read/write logic
Internal: prepare for new build system
- 220.127.116.11 [01-07-2022]
OAuth2 ClientCredentials flow would only work on second try.
dex::PrefixAutoTableWithDatasetto add dataset names in auto-generated table names to prevent potential name clashes
- 18.104.22.168 [31-03-2022]
Conversion errors from string to int/double and int to binary are now passed on to the model instead of skipped.
- 22.214.171.124 [24-03-2022]
Sets in document mappings did ignore
- 126.96.36.199 [23-03-2022]
Labels were right trimmed, but not trimmed from the left.
- 188.8.131.52 [12-03-2022]
Prevent uninitialized warnings during
- 184.108.40.206 [07-02-2022]
Respect the ordering of
name-binds-toindex when writing.
- 220.127.116.11 [02-02-2022]
The maximum line length for CSV files is increased to 64KB.
- 18.104.22.168 [28-01-2022]
Runtime errors within a web service request handler would propagate to a controlling
- 22.214.171.124 [27-01-2022]
Limit Excel sheet names to 32 characters
Allow tables of scalars in AIMMS-generated data sets
Add support, through the
dex::AutoTablePrefix, for auto-generating tables names in AIMMS-generated data sets, based on index occurrence
- 126.96.36.199 [25-01-2022]
Introduced new mapping attribute write-defaults to determine whether for name-binds-to fields, default values will be explicitly written or omitted
Prevent an Excel sheet to be written when it contains no data
Allow write-filter on Excel sheets
- 188.8.131.52 [21-01-2022]
Empty cells in Excel sheet will read to default value, instead of skipping
Empty cells on the first row in Excel sheet will now be skipped, instead of terminating the column range being read
All labels will be right trimmed before adding the a set during read
- 184.108.40.206 [20-01-2022]
Add support for Parquet file format
When constructing a regular expression from the elements retrieved from
name-regex-from, special Regex characters will be escaped.
Regex search for
name-binds-toattributes will take place in a case-insensitive fashion, as set elements in AIMMS are also case-insensitive.
Field names offered for Regex search for a
name-binds-toattribute will first be right trimmed.
- 220.127.116.11 [17-01-2022]
Add support for the OAuth Authorization Code flow for WebUI applications on the PRO/CLoud platform (requires AIMMS 4.84 and PRO/Cloud 2.42)
name-regex-fromattributes for mapping files.
- 18.104.22.168 [02-01-2022]
Refresh token could exceed length of 1024 characters, leading to failed OAuth2 refresh token flow.
Added scope to token request.
- 22.214.171.124 [23-12-2021]
Add support for the OAuth2 Authorization Code and Client Credentials flows to the Data Exchange library. The Authorization Code flow will currently only function on AIMMS desktop sessions. The Client Credentials flow can be used both in desktop and cloud sessions.
- 126.96.36.199 [22-12-2021]
Rows in a CSV and Excel files with an empty value for a binding column would produce duplicate values for the last bound element. - Introduced the attribute
binds-skip-non-existingthat will determine whether to skip rows/objects with an non-existing (or empty) binding or to produce a runtime error.
- 188.8.131.52 [16-11-2021]
dex::ReadAllMappingswould read from a non-existing directory.
- 184.108.40.206 [31-10-2021]
Added support in Excel mappings to map date valued columns to calendars and calendar-valued element parameters.
- 220.127.116.11 [29-10-2021]
Unicode characters taking more than 2 bytes, would not be written correctly to CSV files.
- 18.104.22.168 [22-10-2021]
Introduced new annotation-based JSONDocument generator that creates a mapping for a standardized nested JSON document to read and write all data for a given collection of identifiers in a model.
iterative-resetcan now also specify a list of indices that needs to be reset at a particular node prior to handling all child nodes.
Introduced a new function
dex::ResetMappingDatato empty all identifiers, sets, and reset counters used in a particular mapping.
Changed the default of the
dex::ReadFromFilefunction from 1 to 0, to promote specification-based resetting of counters.
- 22.214.171.124 [13-10-2021]
Allow adding additional suffices to tables in datasets through
Allow specifying custom mapping attributes to identifiers contained in tables in datasets through the
Allow adding row filters for writing tables in datasets through the
Added the function
dex::DeleteMappingto delete previously added mappings. AIMMS would crash when mappings were deleted that contained runtime identifiers from a runtime library that was deleted prior to deleting the mapping.
- 126.96.36.199 [29-09-2021]
The Data Exchange
LibraryInitializationprocedure could crash some models running on the AIMMS Cloud platform
Excel sheets with additional columns without a header in the first row would crash in
- 188.8.131.52 [16-09-2021]
Add support for applying NFC/NFD normalizations to composed Unicode character both contained in the model, or when reading or writing an JSON, XML, CSV or Excel data source.
- 184.108.40.206 [15-09-2021]
When reading CSV files, guess the most likely delimiter
Read/write all values according to the identifier unit/selected convention
dex::ReadMappingsfunction to allow reading mappings from various locations
- 220.127.116.11 [13-09-2021]
Added new function
- 18.104.22.168 [26-07-2021]
dex::ExportStreamContentwould crash for streams bigger than 8 KB
Allow dex::ColumnName annotation to be set on separate index declarations
- 22.214.171.124 [16-07-2021]
Memory streams with binary content could be truncated prematurely when read.
- 126.96.36.199 [14-07-2021]
dex::client::GetResponseHeadersand other functions would not support arguments that are identifier slices.
- 188.8.131.52 [30-06-2021]
Allow memory streams to be read twice by
Allow double values in JSON documents to be read into string parameters
- 184.108.40.206 [28-06-2021]
Add support for memory streams that can be used instead of files in
Add support for dex::client request tracing
Allow reading integer and double values from JSON string properties.
Fixed crash in
dex::client::GetInfoItemswhen calling for string items with no result.
- 220.127.116.11 [23-06-2021]
Support for setting and retrieving headers for up to 4096 characters
Also support GET, PUT and DELETE requests for echo service
- 18.104.22.168 [10-06-2021]
Prevent crash on program exit on Linux
- 22.214.171.124 [28-05-2021]
Updated REST service listener component that used a faulty concurrency setting, potentially leading to connectivity loss
- 126.96.36.199 [26-05-2021]
Added a DLL that was missing in the PROClient IFA on Windows, causing WinUI PRO sessions to fail
- 188.8.131.52 [17-05-2021]
Add a completely asynchronous Curl-based HTTP client to the DataExchange library, supporting all string- and integer-valued options provided by
Add a REST API server to the DataExchange library, allowing model procedures to become available through a REST API via simple model annotations.
ColumnMappingnames to be used in row-based formats such as CSV, Excel, and row- and column-oriented JSON mappings next to the mapping type-specific names available before. This allows for easier switching between various mapping types.
Allow string values up to 8 kB during data transfer with string parameters in the model. The default max string size is 1 kB, which can be changed via the
max-string-sizeattribute for particular string-valued nodes mapped onto AIMMS identifiers.
Add support for transferring sliced AIMMS data via
ExternalBindingmappings that bind indices to the value of an element parameter.
Allow nodes with an
included-mappingattribute to dynamically map the value of bound indices in the outer mapping to externally bound indices in the included mapping. This allows for splitting mappings into smaller constituting components.
Allow an index bound via the
binds-toattribute to become available higher up in a JSON/XML tree via the
Allow read filtering by skipping all data that cannot be bound to an existing element via the
- 184.108.40.206 [08-02-2021]
Introduce new RowOrientedObjectNode and ColumnOrientedObjectNode for JSON mappings, that are both faster and more compact.
max-string-sizeattribute to allow string parameters to hold strings of up to 8KB (default 1KB).
When mapping from/to JSON, the memory used for storing the JSON object in memory would not be returned to the system.
- 220.127.116.11 [17-08-2020]
The library could crash when writing to a workbook with a duplicate sheet name.
- 18.104.22.168 [12-08-2020]
The library could crash because of using a different version of the
libxl.dll(used to actually read and write to Excel files) than the AimmsXLLibrary.
- 22.214.171.124 [06-10-2020]
Added support for reading from and writing to tables in sheets in Excel workbooks
Added support for automatically generating standard Data Exchange mappings from model annotations
Added new mapping attributes
- 126.96.36.199 [27-07-2020]
Name attributes used at mapping locations where no name is needed for a child element are now warned against when reading a mapping
Name-regexattributes used at mapping locations where no name is needed for a child element now result in an error
Boolean values in a JSON file are now correctly mapped onto integer, double and string parameters. During a write the value will be output according to the AIMMS storage type.
- 188.8.131.52 [23-07-2020]
Changed name of
force-denseto indicate that attribute is not only used during write.
- 184.108.40.206 [21-07-2020]
Upgraded internally used library because of performance issue
- 220.127.116.11 [10-07-2020]
Initial public release of the DataExchange library