Release Notes Archive

Studio Runner now has the option to export your generated narrative to a Microsoft Word document. You can export to a new document or use an existing document or template with a bookmark that defines where to insert the narrative. We have released this as a beta feature, and we welcome feedback. See Studio Runner > Exporting to Word.

We have added a Copy to Clipboard button to the Studio Runner interface, making it easier to copy the generated narrative.

The pop-up dialog for selecting your data file was filtered to show JSON files only. We have updated this to show CSV files too, as Studio Runner accepts CSV files for table-based projects.

We've improved the Compose view by adding the option to duplicate a script. Either select a script and click the Duplicate button at the top of the scripts panel, or click the script’s triple-dot icon and select Duplicate from the menu.

We have changed the Function Builder dialog to provide quick links to our reference documentation. Clicking the question mark icon near the top of the dialog now takes you to the documentation for the selected function.

The Function Builder menu is now wide enough to display the full name of each function.

We have re-ordered the menu displayed when you right-click a folder while using the Project Portal in List Mode. Specifically, we've moved the New Project from Sample option up one level to make it quicker to access.

If your script contains a call to a function that takes a lambda expression — e.g. map or filter — and the lambda expression triggers a warning message, that warning now includes the script name, which it didn't before.

We've added a Feedback button to the bottom bar in Studio so you can easily send us ideas and requests for improvements.

We have fixed a bug that prevented the correct re-ordering of scripts after a script was edited. This problem occurred when the sort order for the scripts panel was set to Last Modified.

If you selected a locale from the parameter dropdown in the Function Builder and then selected a different locale, the text identifying the country and language was removed from the previously-selected locale in the dropdown list. This no longer occurs.

Selecting the Move To option in the Project Portal opens a navigable list of folders. Previously, this list could extend off the top or bottom of the screen, making some folders inaccessible. We have fixed this by making the list scrollable.

We have fixed a bug that caused the list of available functions in the Function Builder to be inadvertently cached, which resulted in some functions showing in the wrong project type.

If a collaborator tries to open a project that another user already has open, Studio should issue a warning message and return the collaborator to the Project Portal. This wasn't working because of a bug, which we have now fixed.

If the API endpoint is called with a request body file containing an invalid data section, it now fails with an appropriate error message. Previously, it generated a narrative using the project sample data.

Previously, a #value script containing a syntax error returned an internal server error when previewed. We've corrected this to ensure that the script returns a syntax warning instead.

This release enhances the anomalyDetectorContextual function. The result object for each anomaly now includes fields giving information about the change from the value prior to the anomaly.

We have fixed an issue with the anomalyDetectorContextual function. The function did not perform count aggregation correctly when a group (e.g. Q1) within the selected granularity period (e.g. quarterly) contained just one value. It now aggregates correctly.

pearsonCorrelation returns the Pearson correlation coefficient for two series of numbers.

rollingCorrelation performs a rolling correlation analysis for two series of numbers. A window is rolled over both series, position by position, and each segment under the window is analyzed to calculate its Pearson correlation coefficient.

anomalyDetectorOutliers uses Grubbs' test to detect anomalies (statistical outliers) in a series of numbers. You might use this function to find the high and low outliers in a series of sales figures.

anomalyDetectorContextual detects anomalies in a seasonal context using the selected level of time granularity. You might use this function to detect unusually high average sales for June or unusually low total profit for Q4.

anomalyDetectorExtremeValues detects subsets of contiguous values that are anomalous because all values in the subset are extremely high or extremely low. You might use this function to identify periods in which all monthly sales totals are extremely high.

The isNumber function now has an optional ACCEPT CURRENCY parameter. You can use this to control whether the function recognizes currency strings (e.g. '$500' or 'USD 500') as numbers.

The align attribute is now allowed for HTML tags in Studio.

We have made some minor security improvements.

We have fixed an issue that sometimes caused login authentication to the Studio documentation site to fail.

If a column name in your CSV data contained a dot, the data values for that column would not appear in your project's Data view. This has been fixed.

The groupByJson function takes integer numbers as the inputs to its second and third parameters. If the number was formatted with a pattern giving it a decimal place (or was parsed from a string with a decimal place), the function would fail with a major warning. This no longer occurs.

The groupBy functions no longer produce an erroneous error message when the selected aggregation type is mean() and all values in the aggregation column/field are null.

The initCap function would fail when the input string began with an opening round bracket, a closing round bracket, or a backslash character. The function now succeeds with the initial character unchanged.

Slicing a string element of a list produced by using flattenLists on a JSON array containing arrays of strings would trigger an internal server error. This has been fixed.

If you used flattenLists on a JSON array of arrays, the function would return any string elements with quote marks around them. This no longer occurs.

The createAtlObject function returns an error message when you input a key value of the wrong type. We have changed this message to make it clearer and more helpful.

The NEW PASSWORD field displays a warning message when your password does not contain a number. Confusingly, this message had “[Object OBJECT]” appended to it. We have removed that part of the message.

It was possible to type a non-matching password in the NEW PASSWORD (AGAIN) field but still proceed and change the password. This is no longer possible.

createAtlObject returns an ATL object created from the given keys and values.

addToAtlObject adds key–value pairs to an existing ATL object.

flattenLists takes a list of lists and concatenates the inner lists into one list.

getRandomInteger returns a random integer from between the given upper and lower bounds.

getRandomNumber returns a random decimal number from between the given upper and lower bounds.

You can now copy your project, update the copy separately, then import the updated version back into the original project. This allows you to update a project from a copy while retaining the original project’s API key and endpoint. See Settings > Versioning.

ATL now supports a Python-like syntactic method for slicing lists and strings. You can use this to create sublists and substrings, and to update lists. See ATL Guide > Slicing.

The inflectNoun function now handles most multi-word nouns.

The numToWords function now works on all formatted numbers, not just those formatted using an English-language locale.

The isNumber and asNumber functions now recognize currency expressions with a valid currency symbol (e.g. $) or an ISO currency code (e.g. USD) before or after the number.

The asNumber function now has an optional LOCALE parameter, allowing it to parse numeric strings that don't follow the English-language convention of using a period for the decimal separator and a comma for any thousand separators.

We have enhanced the chooseAtRandom function to prevent it from repeating the most recently used option. The function now has a 'dontRepeat' option, which you can select (or deselect) in your project settings.

Using an ISO-4217 currency code in a call to the currencyFormat function produced an internal server error. We have fixed this to produce a major warning instead. ISO-4217 codes are no longer in use.

If you typed a variable in full ATL — e.g. [[Country]] — while working in Marked Up View, the editor did not automatically add green underlining to highlight it as a variable.

We have fixed an issue with selecting Delete from a script’s three-dots context menu. The Confirm Delete modal appeared but clicking the OK button didn't delete the project.

If your script contained two lines separated by a line break, any space typed at the end of the first line immediately disappeared. This problem was introduced by a Chrome update.

Deleting a project snapshot and then creating a new one caused the deleted snapshot to reappear. This no longer occurs.

With most ATL functions that take a string parameter, the function should coerce any inputted non-string value to a string. However, with ATL objects, this was generating an internal object reference ID rather than a string form of the object.

Studio produced a “Something Went Wrong” error in scenarios where the “Issue marking up your script” warning was required instead. We have fixed this to return the correct warning.

We have fixed a script panel defect that caused closed folders to open when you created a #value script in another folder.

The inclusion of a double quote mark in a column name could cause syntax errors, even when there were no errors in the script. This issue affected "Describe Row in Context" projects only.

With private-cloud and customer-hosted installations, you can set an API data size limit for each project type as well as an overall maximum limit. Setting the project limit above the overall limit produced a “Payload too large” error, even when the file size was below both limits. We have fixed this to ensure the lesser of the two limits applies.

When Studio is used to convert a CSV data file to the API request format, certain characters are HTML encoded. We have fixed an issue where the encoded flag was not set to true.

When you complete an interactive NLG Assist tutorial, Studio should automatically delete the project from your project portal. Instead, it displayed a prompt to delete the project.

The NLG Assist tutorial “Using Language Functions” contained an unclear and misplaced instruction for changing the editor’s paste mode. We have updated the tutorial to correct this.

This release also includes security improvements. The following effects may be noticeable to users:

We have fixed an issue that occurred when directly previewing a user-defined function. When previewed directly, the script's parameters should have a value of null, but instead the first parameter was being given runtime evaluation options as a value.

The filterRows function allows you to apply a filter condition that refers to column names in your table. When a given column name contained a character that must be encoded internally (e.g. &, >, or /), the filter operation was failing with a message stating it could not find the column. This no longer occurs.

We have corrected two issues with the asJson function, both of which occurred when converting an ATL object to a JSON object. First, it was failing to convert an object value that was itself a list of objects, instead producing an object reference string. Second, when used to convert an object that was a list of string values, it was returning a list in narrative form rather than an array of strings.

We have added a whitespace collapsing option to the postprocessing settings. When this option is enabled, Studio automatically removes duplicate whitespace from your project's output text. For example, if there are multiple spaces between two words, the postprocessing collapses them to a single space. It also removes any spaces inserted before a comma.

The currencyFormat function now returns a localized currency symbol when the selected currency is different from the locale currency but uses the same symbol. For example, it returns CA$ for the Canadian dollar when your locale is en_US (English, United States). This removes any ambiguity caused by $ also being the symbol for the US dollar.

Previously, the autocomplete list in Code View returned a "no autocomplete matches" message when there were no matches for your typed text. Following feedback, we have removed the message.

We have made architectural changes to improve how NLG Apps (used via our integrations) are deployed and updated in a customer-hosted or private-cloud environment.

When narrative generation produces too many warnings, the process is stopped with an error message. Previously, when running a "Describe Each Row" or "Describe Row in Context" project (both generate a narrative for each row), the whole report generation would fail with a misleading message about a circular reference. This no longer occurs. Instead, Studio continues with the narrative generation for the next row.

Further to the above, the count of warnings before narrative generation is stopped was being accumulated over the whole run of the project. This has been corrected to ensure the count is performed separately for each row.

If a user-defined function (UDF) returned a string that was then fed through another UDF and then input to the joinStrings function, any commas from the original string were lost. This no longer occurs.

The code is now displayed on multiple lines rather than a single line.

Each section of code can be collapsed or expanded by clicking the arrow beside it.

You can now toggle between the usual light theme and a dark theme, which many people find easier for reading code.

There is syntax highlighting for bracket matching. This helps you detect and fix bracket-matching errors in your ATL or avoid making them in the first place.

You can now quickly locate and replace text using a Find and Replace tool. The keyboard shortcuts (e.g. Ctrl + F for Find) have been added to the Hot Keys list in the user menu.

An autocomplete list of functions and variables now appears when you type [[ or hit Ctrl + Space. The pop-up list filters as you type and allows you to quickly insert function and variable names into your script.

Studio now supports a greater number of locales and currency types. This expands the scope of numerous datetime and number formatting functions. For example, the currencyFormat function now offers a significantly wider choice of currencies.

We have corrected a problem with the groupByVarianceTable function that led to incorrect results when the two input datasets contained different dimension values.

Previously, when you rolled back to a snapshot, then deleted the snapshot, then tried to edit the script, you would wrongly get an error message and you would have to exit and re-enter the project to get past it. This no longer occurs.

When the result from the columnNames or rowNames function is included in the output narrative, it should be printed as a text list and follow the project’s List Formatting settings. We have fixed an issue where the settings were being ignored.

We have fixed the issue where a conditional statement comparing a table cell’s content against a string would not identify a match if the cell value included a forward slash, e.g. “Profit/Loss”.

When a script had two consecutive lines of ATL that were commented out and separated by a line break (<br> tag) rather than a paragraph break (<p> tag), the second commented-out line was still being executed. This problem has been fixed.

We have updated some parameter descriptions in the Function Builder to make them clearer and more accurate.

Previously, changes made in Code View were not saved if you previewed the script before exiting Code View. This has been fixed. You can now preview in Code View and see the changes you have just made

You can now load up to 1MB of sample data into your project.

The Data view for JSON projects now has a Format JSON button. The data editor now formats your data in only two instances: when you click the new button, or when you click outside the editor pane. This makes the editor faster and prevents annoying auto-formatting while you're editing the data.

The Number and Currency settings have been expanded to include settings for the percentage, percentageChange, percentageChangeAcrossSeries functions. These allow you to set project-level defaults for the functions' optional parameters.

We have upgraded our underlying backend software to achieve improved performance, and to allow easier deployment for customers with dedicated systems.

We have fixed the issue where a short-form conditional statement wasn’t recognizing the end of conditional block if the last statement for the last condition didn’t have a terminating semicolon after it.

We have corrected the issue that caused some special characters to be wrongly encoded to HTML entities in the API output.

Previously, the Quick Start tutorial for “Describe Each Row” projects didn’t update the display menu to show the last step (publishing the project) had been completed. This has been fixed.

It is now possible to use the style attribute with the following table elements: <table>, <tr>, <th>, <td>, <thead>, <tfoot>, and <tbody>. This allows greater flexibility when styling HTML tables.

An extra (optional) parameter has been added to the currencyFormat function. This allows you to indicate that negative currency values should be colored red in the project output. A corresponding project setting has been added to the Number and Currency settings page.

The most used locales, en_US and en_GB, have been moved to the top of the locales drop-down list offered in the Function Builder. The same has been done to the locales drop-down on the Number and Currency settings page.

Some performance improvements have been made with regard to opening projects and moving to the Variables view.

Some parameter descriptions in the Function Builder have been updated to make them clearer and more accurate.

The timeout threshold has been increased for running published projects inside Studio with larger sample datasets (on the Publish > Run Project page).

We have fixed the issue that caused a list of numbers assigned to a user-defined variable to be treated as a list of strings.

We have fixed the issue where the expected changes from reverting to a previous snapshot were not visible until you exited the project and then re-opened it. Now, the changes from reverting to a previous snapshot should be immediately visible.

We have fixed the issue where deletion of a script folder (containing scripts) could result in a remaining script having its content replaced by the content of one the just-deleted scripts.

Previously, the 'last updated' time for each script would reflect the last edit in the project (i.e. the last edit in any script). This has been corrected so the 'last updated' time now reflects the last edit for each script individually.

We have fixed the misalignment of the project type icon in the top left corner of the user interface.

There was an issue with using a row name to retrieve rows or cells when the row name was a numerical value greater than 999. This has been fixed.

Previously, when adding ATL to the data source field for a user-defined variable (in the Variables view), any less than, greater than, and single quote characters were being encoded to HTML entities. This issue has been fixed.

A default number format pattern that controls the default formatting of all numbers in your project. This is initially set to display numbers to two decimal places with trailing zeros stripped, but you can change the pattern to your liking in the project settings.

A default locale, which is also an adjustable project setting. Use this to ensure the numbers in your project are, by default, formatted using the appropriate thousand and decimal separators for your region.

Four new formatting functions that allow you to override the project's number and currency settings and apply different formatting to individual numbers:

Project settings are included for each new function. These allow you to set the default values for each function's optional parameters, which reduces the number of parameters you need to specify in individual function calls. In addition, project settings have been included for the pre-existing round function.

Previously published projects are given settings that ensure number formatting works as it did prior to this release. This prevents the new features causing unwanted formatting changes in your published projects. The settings can be changed.

A Deleted Projects view from where deleted projects can be restored or deleted permanently.

A search field for quickly locating projects and folders by name.

The ability to import a project directly into a folder using drag and drop or the folder's right-click menu.

The ability to move a project or folder using its right-click menu.

The ability to select multiple projects or folders using Shift-select.

The portal now remembers which folders you had open or closed (in list mode).

Improved ability to drag a project to a location beyond the top of the visible screen.

Projects with uploaded data now open in the Compose view.

When running a project in TEXT mode rather than HTML mode, it no longer breaks lines by default after 72 characters, though you can still set a line length if you wish. This applies to calling the project by API or from Studio Runner.

ATL objects now allow bracket notation, as well as dot notation, to reference key names. Importantly, this now allows dynamic key access using key names defined in variables.

The contains and indexOf functions now accept 'null' as a search item parameter.

We have fixed an issue with the titleCase function not capitalizing the first word in a quoted string.

For some errors generated while running a project, the API was erroneously returning "Project contains a circular reference: Main -> Main". It now returns the correct error message.

We have corrected inconsistent behaviors in the indexOf function so that failure to find an item in the input list or string now returns -1 instead of an error.

A script interpretation error in the Function Builder's preview mode no longer triggers a 'Something Went Wrong' error. It now produces the correct error message.

Script evaluation errors and syntax errors in the script preview no longer trigger a 'Something Went Wrong' error. Instead, the correct error message is produced.

The groupBy functions no longer ignore nulls in aggregation columns.

There was a minor problem with the Function Builder losing focus when scripts were sorted in a non-custom order. This has been fixed.

We have corrected the defect in the Function Builder that prevented the definition of a call to the concat function from being saved.

We have fixed the issue in the portal where subfolders in a folder sometimes couldn't be seen after being moved there.

Adding script folders without renaming them could result in folders with duplicate names, which then triggered a 'Something Went Wrong' error. This problem has been resolved.

We have amended the hasKey function to be case-sensitive because keys in both JSON and ATL objects are case-sensitive.

The cursor now stays in the script search field after you clear the search term.

We have fixed an issue with "Describe Each Row" projects where user-defined variables were not being recalculated for each row when the project was run with new data.

If a preview warning message included < or >, then the message was being truncated after it. This has been fixed.

We have fixed the issue in the script editor's Marked Up View that sometimes led to characters being dropped.

For various DateTime functions, we have corrected a typo in the Function Builder's description of the locale parameter.

There was a Conditional Builder defect that sometimes meant having a complex condition including both a 'less than' and 'greater than' comparison resulted in part of the condition being stripped out. This has been fixed.

An issue that caused loss of a script folder after changing the sort order.

An issue that caused loss of script folders when exiting a project with script filtering in place.

An issue with the bottom of the Project Portal listing being hidden behind the footer bar.

An issue with the login page if users are logged out of Studio by their computer going into sleep mode or by their Chrome tab being frozen.

Organize your projects into folders.

Switch between the traditional tile view and a more compact list view that shows all folders and projects at once.

Import project files using drag and drop.

Use filtered views of your projects (including a Recent view).

Sort your projects and folders alphabetically or in order of last update.

Larger sample data files (up to 300KB) can be uploaded to a project, and previous limits on the number of rows and columns have been removed.

Studio Runner can now run projects with data files up to 6MB.

The addToDateTime and subtractFromDateTime functions now accept durations in units other than seconds. This includes durations containing multiple units (e.g. "1 year, 2 months and 3 days").

Data columns in tabular data passed to the API call can now be in a different order than in the sample data (apart from the first one).

You can now sort your scripts by creation or modification date.

You can now create a script call by dragging and dropping a script from the scripts list directly into another script.

There are significant performance improvements when running published projects through the API.

We have improved the validation for the Content Line Length field in the Advanced Settings for Studio Runner.

The "Describe Each Row in Context" project type has been renamed "Describe Row in Context" to match the relevant sample project and reduce possible confusion with the "Describe Each Row" project type.

The Preview mode should now only show "Too Many Errors" when there are a hundred or more errors.

We have fixed a problem with repeating snapshot numbers when more than a hundred snapshots exist.

We have corrected a timing issue with autosave that sometimes led to loss of script content.

Previously, rows returned by the filterRows function had columns in which all values were missing (i.e. these columns were being omitted from the result). This problem has been fixed.

Previously, saving a "Describe a JSON Object" project that contained a dot inside a JSON key name triggered a repeating loop of errors. This has been corrected; now the looping no longer occurs and a helpful error message is displayed.

We have fixed a small issue in Marked Up View that made it possible to insert the cursor into the middle of a marked up variable or script name by clicking immediately below it.

The Function Builder no longer tries to preview before the function parameters have been entered.

The Function Builder no longer strips leading whitespace from parameter values.

Previously, subtracting a number from a time (not a date and time) triggered an internal service error. This bug has been fixed.

A minor correction has been made to the information message on the data upload screen for the Convert Data utility (on the Publish > API page).

We have fixed a validation issue in the addToDateTime and subtractFromDateTime functions for non-numeric durations, as well as an issue when the duration was a negative number of seconds.

We have replaced a missing icon in the script editor's color picker.

On a page with more than one copy-to-clipboard button, clicking the second button would copy the correct text but also highlight the text for the first button. This has been corrected; now only the text for the clicked button is highlighted.

Copying and pasting text containing HTML symbols such as > or < or & into another editor such as Notepad no longer results in those symbols being replaced with the encoded HTML entities (e.g. &gt; instead of >). The copied text is now reproduced verbatim.

When you select a word at the start of a line separated from the previous line by a break rather than a paragraph, typing over the selected word now works correctly. Previously, the cursor jumped to the start of the script.

When you enter a validation code after requesting a password reset, Chrome should no longer try to autofill the code entry box with your email address.

The new groupByJsonObject function allows grouping of JSON objects by one or more fields and then aggregation of the values for other fields. The output is the same as for the other groupBy functions, but it takes input data in the form of an array of objects rather than a table or an array of arrays.

The new sublist function takes a list, table row, or table column of data and returns a sublist of the input data. The sublist’s start index and end index points are defined using the second and third parameters.

The "Available Snapshots" section (in the Versioning tab within Settings) now displays snapshot times in the user's local time.

Syntax errors in unused scripts or variables are now displayed when previewing the Main script. This helps users to investigate the syntax errors that are preventing project publication.

When searching backward for a sublist in a list, the indexOf function no longer stops searching when it finds the first instance of the first member of the sublist leading to false negatives.

With the exception of mode, all functions with the "Ignore Nulls" parameter now treat an empty string in the same way as null: if "Ignore Nulls" is set to false, an empty string will be counted as an instance of 0 appearing in the input list.

Comparing a number to an empty string no longer leads to an internal service error. Now, such a comparison returns false, as a number can never be equal to an empty string.

Previously, some scripts disappeared when multiple scripts were dragged from one folder to another within a script menu already sorted in ascending alphabetical order. This issue has been fixed.

The Insert Script drop-down list in the editor once again displays scripts at every level of the folder hierarchy rather than just the first two.

Previously, styling was not always applied to the first item in a script when that first item was a piece of marked-up ATL (such as a function or variable). This issue has been fixed.

Dragging the mouse outside of the Conditional Builder modal no longer closes the builder.

The Conditional Builder now shows the appropriate warning when incompatible types are compared to each other.

The Conditional Builder's focus no longer jumps back to Parameter A of Case 1 when additional cases are added.

The Function Builder and Conditional Builder can now be used in full-screen mode.

A more helpful error message is displayed when you attempt to publish a project with a user-defined function not matching its script name.

Previously, enabling automatic sentence capitalization (Settings) sometimes led to an internal server error. This issue has been fixed.

After entering invalid settings on the application configuration screen in Studio Runner and then clicking off the informational modal, the main Continue button no longer continues to display a spinner.

New math functions

New dateTime functions

New editor features

Dragging the mouse outside of the Function Builder modal no longer closes the builder.

The Conditional Builder no longer attempts to produce a preview until all necessary fields are filled in.

Using the editor's Clear Formatting button in Marked Up View no longer erroneously removes ATL brackets.

Text alignment inserted using the editor’s styling buttons is now saved correctly. Note that this will not be reflected in an API output if the “text” option chosen is contentOutputFormat.

The "Open in new tab" option of the Insert Link dialog is now correctly stored in the HTML tag.

Copying and pasting no longer turns <br> tags into <p> tags.

Clicking on a top-level category in the Insert Function drop-down menu no longer closes the submenu for that category.

The script list no longer scrolls to the top when a script other than the one currently focused on is deleted.

A number of small UI issues have been cleaned up.

The Data view

The Variables view

Function Builder and Conditional Builder

Functions and ATL

Editor

Studio Runner

Customer-hosted deployment

Miscellaneous

Additional functionality around script management:

Additional functionality around dates and times:

Many new ATL functions have been added to Studio:

Improvements to Studio Runner:

Improvements to privately hosted deployments of NLG Studio:

Syntactic errors are now shown along with other warnings in the preview pane rather than preventing preview altogether. However, projects containing syntax errors can no longer be published (this is to avoid users being charged for running them).

The decodeHtml function has been improved to allow manual insertion of HTML <span> tags with a style attribute, for example making it possible to color small chunks of text, change the font, or add margins.

On pop-up dialogs, pressing the Enter key now results in the same action as clicking the dialog's main button.

The group of sample projects available directly within Studio has been extended, and some minor updates have been made to the Table Fundamentals and JSON Fundamentals projects.

The warning and error messages around illegal script and folder names have been improved.

The list of illegal names for variables has been updated.

Studio now tells users when a new version is available with a prompt to refresh the browser to avoid glitches when an old frontend version is used with a new backend.

The Convert data service, on the Publish > API page, now accepts data files up to 2,000,000 bytes.

It is now possible again to create a new variable directly from the parameter drop-down menu in the Function Builder and Conditional Builder.

A bug preventing changed variable definitions from being saved has been fixed.

A bug preventing the Conditional Builder from building or previewing conditionals using the isEmpty test has been fixed.

A bug preventing single variables or function calls being copied and pasted in mark-up mode has been fixed.

A problem with the filtering algorithm has been fixed making sure that matching items inside collapsed folders also get found.

Attempts to convert sample data with duplicate row or column names to API format now receive the appropriate error message.

Some error messages in Studio Runner have been improved.

A bug preventing the command sequence Ctrl/Cmd-A Ctrl/Cmd-V (i.e. overwriting the complete content of a script) from being undone has been fixed.

An issue preventing one user from opening a project that was already being edited by a collaborator has been fixed.

An issue in JSON projects has been resolved, whereby previously, saving a variable pointing to a leaf value reset the type to JSON instead of string or numeric.

Cancelling the regeneration or deactivation of a project’s API key no longer causes an unexpected error.

Editing JSON data such that it exceeds the size limit and then cancelling the warning dialog no longer leads to a “Something went wrong” error.

Selecting a piece of text then previewing it using the preview hot-key combination could result in the selected text being deleted. This has now been corrected.

Occasional problems with copying and pasting marked-up content in the editor have been fixed.

An issue with the privately hosted version of Studio intermittently showing wrong links on the login page has been corrected.