2nd MAY 2019
Feedback Tasks: We have released completely new functionality to facilitate the feedback process on data sets in order to train the Onedot software! Onedot now offers the concept of Feedback Tasks, a light-weight mechanism to get feedback from one or several persons collaborating and answering questions raised by the Onedot software in case of low confidence, missing context or simple ignorance of some product ranges the software has never seen before.
You will now see a new icon in the dark-blue navigation bar, clicking on it will open the new task view. Here, you can see all tasks assigned to you, where your feedback is needed. Double-clicking the task will open a data set view where you can submit your feedback.
Follow the instructions provided in the column headers that describe where and how your feedback is needed. Once you are done entering your feedback, click on Resolve to resolve the task and send it back to Onedot. Processing will then continue and take your feedback into account for the current processing step.
Availability of Checkbox Column: The checkbox column functionality is now also available when there is only 1 initial distinct value in the column selected. For this to happen, we are now also allowing to map the checked/unchecked state to arbitrary values (e.g. values which do not exist yet (and are therefore not mapped) at this point in time in the checkbox column).
Conditional Allowed Values: A new specific type of allowed values is now supported as well: Conditional Values. This functionality comes in handy if you want to use allowed values which depend on a certain condition, e.g. attribute values which are only valid for certain product categories.
Simply select two columns where the first one is the condition, e.g. the product category, and the second one denotes the allowed values column, e.g. the category-specific attribute value, and right-click to choose Conditional Values. You will be asked to select a data file containing at least two columns (the condition and the allowed value column) or alternatively a dictionary file.
Pressing OK will set up the allowed values in such a way that you can only choose from a subset of values, depending on the selection of the condition column.
The conditional value is its expected value or the value it would take on average over an arbitrarily large number of occurrences, given that a certain set of conditions is known to occur.
Better Behavior of Checkbox Column and new Rows: You cannot check the value of a new row in a checkbox column anymore, avoiding all kinds of inconsistencies and unexpected behaviour. Luckily, no more travelling to weird places for our users.
24th APRIL 2019
Derived Values: A new type of column has been introduced in the Onedot Application: Derived Values. This type of column shows values which are uniquely depending on some source values, and are updated in real-time when a source value changes.
Use cases for this new column type include:
- Automatically showing a (conditional) histogram for a target attribute in an attribute mapping report. When you change the target of a mapping, the histogram changes as well, allowing you to immediately verify your decision.
- Supporting key/value-based attributes, category tree levels, etc. Usually, enterprise-grade product data models (PDMs) use business keys to identify concepts like attributes and categories in a language-independent way. Derived value columns help to keep business keys and related information in sync while you are editing product data sets.
- Using dictionaries to transform values in real-time: You can use a dictionary as the source for the derived values, which allows you to implement use cases like brand name normalisation on-the-fly.
This new functionality complements the existing Allowed Values and supports more sophisticated use cases of data validation and dependencies.
Task Workflow Conditions: We have introduced additional checks to perform when resolving tasks. Onedot can automatically verify whether all feedback columns have been filled or whether all feedback columns with allowed values have only valid values specified.
Clicking Resolve on a feedback task will trigger these checks and show appropriate error messages in the Onedot application if not satisfied.
Fixed Horizontal Scrolling: An issue related to scrolling horizontally in the data set view made life quite hard for some of us. This in particular happened using late Chrome browsers.
While using the Tab key is a valid workaround, life without scrolling nowadays is just unthinkable. Luckily, you can use your mouse wheel and scroll bars again and enjoy the usual behaviour you are so used to.
17th APRIL 2019
Better Reliable Save: When performing a series of actions on a data set, e.g. filtering, sorting and editing, all at once, sometimes we were not able to save these actions appropriately, as one would expect. We have identified and solved the problem so evil strikes no more.
Dataset View features Checkbox Columns: While editing a data set, e.g. as part of a feedback task, it is paramount that you do not loose the focus when giving feedback to some hundreds of products or product groups where the software came up with low confidence results.
That’s why we’ve introduced a new type of column called a Checkbox column (in addition to Regular columns supported until now). Checkbox columns are optimised to track feedback status in a data set, e.g. whether a record has already been processed during feedback (either confirmed as being correct, or edited to correct the software’s decision).
The Checkbox column can be toggled simply clicking on the relevant cell. Even better, if you correct a record by editing a cell, the Onedot Application automatically records that feedback has been given to this particular record and checks the cell without any further interactions, signalling that feedback is done for this record.
Asking for Save: Saving huge data set is generally not possible if it exceeds a certain limit of cells. Despite that, the Web browser still asked user to save the data set in such cases, causing trouble. Rest assured, it does not do that anymore.
Nicer Visuals for Navigation Bar: The left navigation bar of the Onedot Application has been visually improved, in particular if expanded.
10th APRIL 2019
Automated Detection of Field Delimiter for Text Files: The Onedot Application has become a lot smarter when it comes to detecting field delimiter for text files, such as CSV files where you encounter various delimiters such as comma, semicolon, tabulator or other custom characters.
The software analyses the uploaded text field and automatically detects and uses the appropriate field delimiter and character encoding, making text file handling for product data exchange a breeze.
Fixed Glitches while editing Cells: A number of smaller glitches has been fixed involving the editing of data set cells. Overall, we improved the speed and visual presentation, so larger data sets are more pleasant to edit than earlier.
Histogram View Improvements: The previously introduced Histogram view now also has support for allowed values, further simplifying the mass editing of product attribute values in the context of more complex product data models with huge lists of values.
Moreover, the visual representation of the histograms has been improved, making the whole view easier to read in case of larger datasets.
General User Interface Changes: We are constantly fine-tuning our Onedot Application user experience in order to make sure our product is intuitive and pleasant to use.
Several parts of the user interface have been streamlined, aligning terminology, visual representation and behaviour in order to provide a smoother experience for our customers.
3rd APRIL 2019
New Histogram View: Data sets in the Onedot Application have got a brand new view called the Histogram view which is available next to the existing Dataset and Statistics views. The purpose of this new view is to show the histogram of a data set, meaning only the distinct values per attribute, with an indication of how many occurrences of a particular value.
The Histogram view is designed to support bidirectional editing, so if you edit a value in the Histogramview, the changes automatically get applied in the Dataset view as well, and vice versa. Of course, sorting and filtering works as with any other view.
This new functionality is particularly useful when profiling product data or performing manual mass-editing of large product data sets.
More Meaningful Statistics: The Statistics view of a data set now shows a warning stating that statistics are only taking a sample of huge datasets into account. Previously, we where just showing the results based on the sample, causing unnecessary confusion. Now, we show it loud and clear, for your safety.
Filtering for Blank/Filled Cells: You can now also filter for any blank values in a column in the dataset view by just typing (Blank) into the filter text field. Similarly, if you type (Filled) only rows where the column has a value filled in will be shown.
Sorting has now three States: Previously, sorting in the data set view only supported two sort orders, ascending and descending. Now, there is a third one, using the order of rows as in the original data set.
Prefix Matching for Filtering and Suggestions: Typing Space followed by the text you want to filter for now automatically matches based on a text prefix. As an example, consider the following:
- Typing just green in a column with the colours green, light-green and dark-green would match all colours mentioned.
- Now, typing Space followed by green (meaning “ green”) would only match green, but not light-green and dark-green.
The same also works in columns with allowed values: suggestions are also filtered based on this new behaviour.
The new prefix matching functionality is particularly useful for semi-structured text like product descriptions, category tree labels, etc.
Quicker Display of Sparse Datasets: The infrastructure to read data sets and display them in the Onedot Application has been improved with respect to speed and reliability. Larger data sets whose columns are only sparsely populated with attribute values now display a lot quicker, in particular with low-bandwidth network connection.
Pressing Escape resets Value: When editing a cell with allowed values configured, pressing Esc when suggestions are shown will not only close the suggestion window, but also reset any previously typed text, e.g. when searching for appropriate allowed values. This facilitates categorisation and schema mapping feedback when exploring possible options for category or target attributes.
27th MARCH 2019
Sorting/Filtering per User: We have made some improvements to the stability and reliability of storing sorting and filtering settings of a dataset individually per user, no matter which browser or which computer is being used. When you leave your browser with a dataset displayed and reopen it again the next morning, you’ll find it exactly as you left. Even better, two persons collaborating on the same dataset can use different sorting and filtering settings!
Show Columns Action always there: Previously, you could end up in a really dark place if you did hide all columns of a data set, there was no way of showing the columns again.
The Show Columns action is now always available, helping you to escape this uncomfortable situation and show all columns again.
Stability Improvements for Hierarchical Allowed Values: Using Esc key in combination with hierarchical allowed values works now as expected: You can choose allowed values from the drop-down, type any text to get suggestions on possible allowed values, press Esc to revert back, etc. This provides more flexibility when the exact to be used are not known in advance and a little exploration is called for.
13th MARCH 2019
Tooltip for large Values in Dataset View: We have added a tooltip in case the cell value is not fully visible in a dataset cell, which is frequently the case for product descriptions and other free text attributes. Smoothly hover on the cell in question, and a tooltip is shown after a short delay, showing the entire cell value at once.
Loading Visualisation: The animation shown while loading a dataset is now way prettier than before on certain browsers. The dots now smoothly beat while the data is processed. Problems related to clipped fonts and animated dots are a thing of the past.
Fixed Dictionary Validation: Detection of chains in dictionaries is working better even for very non-obvious cases. The dictionary view shows any chains, duplicates and forks as one would expect.
Chains are usually a problem when repeatedly applying a dictionary on the same dataset again, causing the results to change each time, which is not wanted in all cases.
20th MARCH 2019
Smarter Focus Handling: Some dialogs such as the Allowed Values dialog did not behave very friendly in terms of setting the input focus when you need to type something. Now they behave like good citizens again, accepting your input from the start without having to click into the input text field anymore.
Fixed Crash adding Row/Column: A deadly combination of creating a new column on a dataset after adding the new row previously led to a full stop of the application, requiring a refresh of the browser. Thankfully, the patient has been cured and survives such conditions without any problems.
More Beautiful Tooltips: The visual design of any tooltips, e.g. the one appearing on the Download button when attempting to download a dataset without prior saving, is now better aligned with the other parts of the application.
Fixed Behaviour on Enter: The behaviour when pressing Enter after editing a cell has been corrected. Sometimes, in combination with hidden columns, the focus would be set on the wrong cell one line down. Not anymore, you can proceed to edit with full speed.
20th FEBRUARY 2019
Correct Alignment of Decimal Numbers: Decimal numbers such as quantities, e.g. 1.23 cm, are correctly displayed with right alignment, also taking the decimal point into account. This is how it should be and we are very happy with that feature.
Fixed Header Comment Action: The action in the dataset view to click on a column header comment has been fixed and works again as expected.
Sorting/Filter Settings tracked per User: Any sorting and filter settings are now separately tracked for each user. Collaborative editing or feedback on larger datasets has become a reality! On top of that, this also simplifies setting up demo cases which are guaranteed to look how the should look like, regardless of the browser you take.
27th FEBRUARY 2019
Semantics of Sorting/Filtering: Changes to the dataset do not cause the dataset view to behave erratically. When you edit a cell which would be filtered away, it still stays visible, yielding a more stable editing experience. Same goes for sorting.
Better Editing Performance: We made editing performance of the dataset view better, in particular in combination with hierarchical allowed values. Some third-party components were upgraded, the functionality of keeping track of allowed values and suggestions enhanced, making editing larger datasets a blast. Enjoy!
6th FEBRUARY 2019
Double-Click to Fill Value: Double-clicking at the light-blue square on the bottom right corner of a cell selection will automatically fill the value of the selection to the cells in the column, right until existing cell values are encountered. This feature reminds us a lot of: Excel.
Quick Renaming of Files, Folders and Tasks: The details pane of files, folders and feedback tasks now allow renaming in-place. Double-click the name, type the new name, done. Works with descriptions and other text properties.
Improved Performance editing Allowed Values: Bulked up the browser cache and avoiding unnecessary computations for luxurious datasets of 50K rows or more with associated category hierarchies of up to 10K categories. Plus, we reliably fill hierarchical allowed values, even with inconsistent or missing existing cell values.
Improved Allowed Values Drop-Down: The drop-down showing suggestions for allowed values now even works if there are existing, invalid cell values. It was painful to work without suggestions shown in this case, making it hard to fix invalid cell values. And to make it even better, the sort order of the suggestions uses now standard alpha-numeric sorting.
13th FEBRUARY 2019
Fixed Sorting of Folders: Storing the folder sort order has been improved. Previously, it did not survive a refresh in the browser, now it stays around even after the refresh. Good to know and should be used accordingly.
Improved Allowed Values Dialog: The dialog to set allowed values has been overhauled. We now show any values as tags, making it easier to edit more complex values including special characters. As an example, it is now also possible to use allowed values containing commas and other characters.
Row Numbers: The dataset view now shows row numbers for each row in a dataset. This helps to better identify a specific row and also allows you to better see your progress when giving feedback.
Note that row numbers behave slightly different than in Excel. When filtering and sorting rows in a dataset, row numbers will change, while Excel keeps them stable.
“Onedot supports ZAGENO in our rapid market expansion by automating the vendor onboarding and making products more searchable through better product data quality. This ultimately leads to a faster growing global customer base and higher customer satisfaction.”
David Pumberger, CEO & Co-Founder at ZAGENO Inc.