Featured Ideas
Hello,
After used the new "Image Recognition Tool" a few days, I think you could improve it :
> by adding the dimensional constraints in front of each of the pre-trained models,
> by adding a true tool to divide the training data correctly (in order to have an equivalent number of images for each of the labels)
> at least, allow the tool to use black & white images (I wanted to test it on the MNIST, but the tool tells me that it necessarily needs RGB images) ?
Question : do you in the future allow the user to choose between CPU or GPU usage ?
In any case, thank you again for this new tool, it is certainly perfectible, but very simple to use, and I sincerely think that it will allow a greater number of people to understand the many use cases made possible thanks to image recognition.
Thank you again
Kévin VANCAPPEL (France ;-))
Thank you again.
Kévin VANCAPPEL
Vanilla Alteryx Chained Apps can only progress linearly, which means developers could not let users skip few applications ( or ) reach the last app in the chain ( or ) let the user select which specific app to trigger based on the requirement.
This can be bypassed by using a render tool with output as PCXML and HTML link of the Application you can trying to divert to, which does not affect the existing workflow in any way.
By using the below set of tools on any workflow/chained app you can either branch the flow of apps ( or ) you can skip a few apps in the chain.
- A excel or csv file which has the links of the apps - the reason for keeping the hyperlinks in an external file is so that we can update the link if the server link changes/updates - refer Image 1
- A filter tool to specify which application to move to ( can be changed using a radio button/drop down to app 2/3/4/5 etc.)
- A text tool ( This is where the magic is ) - configure it to pick the server link from the incoming data from the filter tool as hyperlink and generate a output preview, as shown below - refer Image 2
- Use a render tool as output and write to any PCXML file ex: "File.pcxml" - refer Image 3
Image 1 - Input Configuration with the flow that can be part of any existing application
Image 2 - Text Tool Configuration
Image 3 - Render tool Configuration
POC in action
- Let assume our 3rd application is located in www.alteryx.com - if the user selects 3rd app in the radio button
- Which would generate an Output Preview like below
Now If clicked on App 1, it would divert me to www.Alteryx.com
Regards,
Maithreyan S
Whether it is a date range or a numeric range, I would like an official tool that can tell me where the gaps exist within the range.
The Find and Replace tool currently replaces text or appends fields by reviewing the first word in the string and evaluating if there is a matching value in the lookup replace table. If there is, it returns the replace text or it continues onto the next word in the string and repeats until a match is found or not found.
However the functionality should be that the entire string is evaluated and the first match in the replace lookup table is returned. This is similar to how the VLOOKUP formula functions in Excel.
This makes workflows easier to maintain as the ordering of the lookup table can be amended to return the appropriate value. The ordering of the source string is far trickier to automate/change.
I have seen workarounds where append fields is used to append all lookup values to each row and then find rows which contain the string and use a sample tool to return the first row, however this causes processing issues when working with large number of records/ lookup values.
The functionality should either be changed or at least a setting which can be toggled to switch between the two options.
Hello,
We would like to have further options in the test tool. As of today you can only test if records matches (equality test) , we would like to be able to test also when it is greater (>) and smaller (<).
Thanks !
Hi all
Currently when you set your workflow to don't write outputs (disable all tools that write output) under runtime of the configuration of workflow- the render and green output tools become greyed out and do not write an output (as expected).
However, this is not the case for connectors - for example, if you use the SharePoint output tool and click disable all tools that write output, it will not be greyed out and still write an output. Is it possible for these connectors to also not run when this is selected in the configuration? As otherwise currently, you have to add it to a container and disable it.
I love the ability to cache data. Sometimes, I've cached too far and I need to go back and alter a tool's configuration or make some other cache-destroying change. I then have to recache somewhere and sometimes the dataset is so huge it is a pain. I'd love to create multiple cache points as a safety or backup.
In conjunction with the "First N% of Rows" and "Group by column (optional)" features, an additional option that would be great to have would be to tell the tool to take at least 1 record per group. If I am telling the tool to take the first 3% of records and grouping by a field, but one of the groups only has 5 records, it won't return any records for that group, as 3% of 5 is .15, so the tool rounds that down to 0. If I could also choose an option for the tool to always round up, that would also be sufficient.
In the formula tool, you can generate uniform random values using the RAND() function.
I would like to have similar functionality to get normal random values by calling RANDN().
As a workaround, the values can be produced from a Python tool using
numpy.random.randn()
Python documentation on numpy.random.randn.
The full script below produces 100 random normal values in a column called 'Normal_Rand'.
The zip file has both a direct workflow and macro solution to produce the random normal values.
Would like to have the ability to filter to searched fields. Having the ability to search for already known fields, and then selecting those fields, temporarily, and isolating them for view. While there is the selection of fields under "X of Y fields", this still requires manual clicks and could be tedious when there could be a large amount of fields.
While you could also use the SELECT tool, that would still require manually searching for the field and isolating it for viewing. This could possibly be an enhancement for the SELECT tool or it could be an enhancement for the "X of Y fields".
I'd like to suggest that Alteryx make annotations able to be hidden or visible by right clicking or hovering over a little highlighted corner similar to comments/notes in Excel. Have a highlighted corner of tools when there is an annotation. When hovering over the colored corner, user is able to view the annotation otherwise it is hidden and out of the way of the tool box.
I know y'all are working on data lineage for some future offering and it is very much needed. For highest quality results, please make logs a primary source of lineage information. Being able to use dynamic naming with some tools and macros makes the names in the workflows simple foobar placeholders and do not reflect what actually happened. Today Connect doesn't use logs and leaves many lineage gaps because of this
Please move this to a more appropriate category if needed. This future feature work is not part of Connect.
Hello all,
My idea is very simple, and this is most likely I'm used to do. After highlighting the fields in the select tool, the next action is to click Options and to apply what to do. My idea is just after highlighting, right click on mouse button to go exactly on Options menu. Just a simple right click. I think this behavior is most likely comes from Microsoft, I'm used to select multiple lines or cells and with right click to do the next action. It's a simple, but powerful. What do you think?
Albert
The DateTime Tool currently displays incoming date-time data types in a dropdown menu and "If no date-time columns are present, the field appears dimmed and shows 'No Suitable Input Fields Available'". However, considering the complexity of date formats, it's beneficial to directly view the data type of incoming date-time data instead of relying solely on visual inspection. While Designer filters options based on string or date/time types automatically, having visibility into the data type ensures a more sound interpretation of the input for analysis purposes.
When making changes to a UI, it would be beneficial if I could move objects horizontally instead of only vertically. Additionally, being able to resize items in the interface designer could improve end user experience. Currently I have List Boxes that are cutting off some words in a workflow and if I could extend the box or wrap the text then I wouldn't lose any context.
The second idea is the freedom to sort the workflows in collections that I control in anyway that I choose.
I pull data from APIs frequently in my workflow, which has some potential for error if the API is unavailable. I would love to be able to develop the workflow without having to add in some text box place holders or record count feature to allow the remainder of the workflow to run without error and potentially be able to try the API again so if "Invalid authentication" route to a section that will regenerate an authentication and retry the data call. I have currently worked out some hacks to be able to work through the issue but it makes the workflow very messy. A tool like the Detour tour that goes if value contains 400 go this way. If value contains 200 then close off this container and continue on the other one. Maybe the request is to allow the detour tool to be set to a specific value and maybe have multiple route options if A then do connection one, if B then do connection two or If C then do connection three.
I think it would be very helpful (and very easy) to add metadata to a yxdb with the name and location of the workflow that created or most recently updated that file. Right now I have to add a column to the data frame with a formula tool with the location, which puts the filepath on every single line, and is not how I want to present a finished product dataset to a superior. Thank you!
Currently, you can only type forward in the Description Window in the Meta Info tab. So once your description is in there for your workflow, you have to retype the information or type it in notepad and save it back in there. It would be nice if you could type on any previous line in that window.
It would be nice to also have some additional fields that appear to be common among the community, such as Data Sources and Purpose. That way you could use the Description window for version notes and any other notes for the user.
The Directory tool should also retrieve the author/creator name of a file along with an audit trail of anyone that made changes to the file. The audit trail option may need to be a different preparation tool but the information would be useful.
Providing detailed information about inputs, outputs, data sources and assumptions provides useful information to us that may not always be obvious within a workflow, especially one with many tools used. While the comment tool is useful, and I enjoy the customization options, I think some updates to the annotation tool would be useful.
One improvement I would like to see is the ability to summarize and download all annotations to a PDF document; color coding the annotations according to the tool used would also be helpful. Having this information readily available for presentation purposes or for enterprise-wide users of workflows would be beneficial.
If this is already possible, please let me know!
The Find Nearest tool allows setting a distance in units of miles and kilometers. It would be incredibly convenient to specify units of feet or meters also. I find that in most instances, 1 mile is vastly large for the analysis I am working on and I have to include additional logic to confirm incorrect matching is not occurring.