This release adds improvements and bug fixes to v1.29.0. The download links are at the bottom of the page.
We recommend that you back-up important transform files before installing any new release. Please let us know of any problems or issues you encounter with the release.
You can uninstall old releases after you have successfully installed this version.
Windows and Mac
1. A new Cluster transform has been added. It classifies a column of text according to the closest fuzzy match to a list of terms.
You can use this to classify items in a dataset:
Or to fix misspellings:
Set Clustering to:
- Manual to cluster by Terms you provide (one per line).
- Automatic to try to find clustering terms automatically.
- Guided to try to find clustering terms automatically, starting with Terms you provide (one per line).
Set Closeness to how close a value has to be to be considered a match. Values that do not reach the minimum closeness to any term are set to cluster value <None>.
For Automatic and Guided clustering:
- Set Max clusters to the maximum number of clustering terms to use.
- Set Max time to the maximum number of time (in seconds) you want to allow.
See this video for more details.
2. The Column Values window now allows you to change how many values to show and whether to show the most or least frequent values. Thanks to Dieter S. for the suggestion.
3. The Average length (the mean number of characters of non-empty values) is now shown in the Summary transform and the Column values window.
4. The number of negative, zero and positive numeric values are now shown in the Summary transform and the Column values window.
5. The Sample drop-down list in the Characters tab and the Column values window is now shown in red if the dataset is being sampled.
6. The Min length value in the Summary transform and the Column values window are now only computed for non-empty values.
7. The Sum numeric value in the Summary transform and the Column values window are now shown as N/A if there are no numeric values.
8. Previously values were passed to the Javascript transform as a Javascript Number, Boolean or String type, depending on the value in each cell. This could be problematic when values in the same column could be interpreted as different types. So all values in a column are now interpreted as the same type. There could also be an issue where values that could be interpreted as numbers were better interpreted as strings (e.g. ID values with leading zeros). So there is now new Type columns as option which allows you to specify whether to convert each column to Javascript Number, Boolean or String type (depending on the contents of the column) or convert them all to Javascript String type. Thanks to Jan V. for reporting this issue.
9. Previously Easy Data Transform would sometimes hang trying to display datasets with 50 millions rows or larger. This was due to the table size exceeding 2^31 pixels! So it now only display datasets up to 10,000,000 rows or 100,000 columns in the Right pane. Thanks to Bert for reporting this.
10. Fixed a crash that sometimes occurred when selecting JSON input items. Thanks to Jindi L. for reporting this.
11. Previously command line processing would continue even if there were issues with the command line arguments, for example:
- -file set to an input file that did not exist
- -file command syntax incorrect.
It now exits without processing. Thanks to Jan V. for reporting this.
Mac only
12. Fixed issues where some controls were not correctly repainted when switching between light and dark themes.
Windows Download
Mac Download