In the Tableau Forums, we frequently request those who post questions to attach a packaged workbook (for Tableau Desktop), or a packaged flow (for Tableau Prep Builder), or a link to a workbook on Tableau Public (for Tableau Public users). This document is to help you understand when a packaged workbook or flow or Tableau Public link helps, why we ask for it, and how to include one. For the rest of this post "workbook" can be interpreted to mean "workbook or flow or Tableau public link". There are four sections:
- I'm not able to attach a packaged workbook, so what can I do next?
Nearly every time. The list would be endless. It is easier to answer when a packaged workbook probably won't help:
- When the problem absolutely cannot be reproduced using an extract, and you have tried and proven that.
- When you are using the Public (free) version of Tableau Desktop, in which case there is no capability in Tableau Public Desktop to export to a packaged workbook. In that case, we ask that you publish your sample workbook to Tableau Public and include a link to it in your post, so we can download it from there, for all the same reasons below.
- Screen captures and descriptions are rarely enough. We need to see a realistic representation of the data structure, fields, data types, calculated fields, placement on various shelves (row, column, color, shape, size, etc.) If "a picture paints a thousand words," then a viz paints a thousand pictures!
- If you ask questions about writing R, PowerQuery, Python, SQL, etc. then you typically share your code to get help. A lot of Tableau Desktop & Tableau Prep's configuration is done via point & click & drag & drop so sharing the workbook is sharing your "code" so you can get the help you need.
- It helps you clarify the problem, for yourself as well as for us. Even for myself, I know I have had many times when I was trying to properly document and build an example, that the solution revealed itself. Questions with examples tend to be more complete, clear, thoroughly considered and better constructed, thus easier to answer.
- It saves you time! Really. You can spend a lot of time trying to explain the problem, the data, the goal, what you have tried, and conversing about a variety of details and requests for a workbook. Or, you can just attach one and we can get down to business.
- It shows that you have made a good-faith attempt to solve the problem yourself. Most of us do this in our spare time and are unpaid. We volunteer our time to help you, and we are happy to do so to help you learn Tableau and overcome challenges. But we are not here to be a free development service, or to do your homework for you. If you can't even get started, then you probably should begin with viewing some of the free on-demand training videos. Poor Richard's Almanack once said, "God helps those who help themselves." Well, we aren't God, but we feel that sentiment sometimes.
- Courtesy. It is common courtesy to cause as little effort as possible to the people who are going out of their way to help you. If a friend asked you for a favor, how would you feel if they made you do all the work while they sat back and waited? That is the position we are in.
- You are more likely to get a response. Many of the volunteers that help here use a tool called the Crow's Nest to find questions to which to respond. Crow's Nest color codes the unanswered questions by the presence of and type of attachments. Frequently, we/they will skip over those without a packaged workbook, for all of the above reasons.
- Our responses will use the same version you use. Imagine how frustrating it is for both you and us to spend time creating a solution for you and posting it, only to find out you can't use it because you aren't on that version.
- Context and possible side-effects. Frequently, after providing the correct answer for the specific question asked, the original poster (OP) follows up with another question, saying something like, "That worked, but now something else broke." That "something else" we didn't know about because it was not included in the question. If we had the context, in the form of a packaged workbook, we would likely have seen the possible side effects, and the answer could have been more complete, saving both of us time.
- Export a packaged workbook (or packaged flow for Tableau Prep). A packaged workbook is a workbook file that has the data packaged into it so others can open and view your viz without having access to your data source or other resources. A regular workbook (.twb) is a text file in an XML format that defines how to construct the view based on the data in the connection and other resources defined in the workbook. On the other hand, a packaged workbook (.twbx) is in a binary format that includes the data itself as well as other objects in the view. For Tableau Prep a regular flow (.tfl) is a binary file that includes the flow, a packaged flow (.tflx) includes the flow and the data.
- See Packaged Workbooks documentation for details.
- Remember, if your data source is a live connection to a database, like SQL, Redshift, etc., you must extract it. Creating a packaged workbook with a live database connection is useless, because we don't have access to that database.
- Attach it to your post. See Attaching a Packaged Workbook.
- Please document the Tableau version you are using. This saves us lots of time trying to open it it in different versions until we stop getting error or warning messages. Please make it easier on us!
- "I have confidential data" is one of the most common reasons given for not including a packaged workbook. If you have confidential data, here are some options for you to create a sample packaged workbook without compromising your confidentiality:
- Create a mock-up using the Tableau Superstore database, the Coffee Chain database, or some other fake data that you made up. For example "Customer ID" in Superstore could be patient id, "Order ID" could be your visit or encounter ID, and "Product Name" could be a particular lab test.
- Replace any confidential information in your data set. See Anonymize your Tableau Package Data for Sharing |Tableau Support Community for a detailed explanation of how to do this.
- Remember that numbers, database identity fields, dollar amounts and even geography like zip code or latitude and longitude, have no meaning in themselves. Things that you should not include, and need obfuscation:
- Bank account or credit card numbers
- Names, addresses, email addresses, phone numbers, birth dates, and names of families or individuals
- Company names, brand names
- Government or military targets or operations
- Remember, while this may take what seems to you like a lot of time, it really takes less time than all of the discussion that occurs trying to avoid mocking up or blinding your data.
- Clearly document your goal. If you don't know your goal, you will never find it.
- What are you trying to understand about the data?
- What questions are you trying to answer?
- How will the answers to those questions be used to make business decisions?
- Include a Truth Table or its equivalent, showing all of your possible input values and your corresponding desired or expected output values. Make sure you have considered each possible variable.
- It is also helpful if you document what resources you have used, so we don't just point you back to the same ones you already saw.
For various reasons I'm not able to create a packaged workbook, so what can I do next?
- Check once more whether that's at all possible. Please trust us when we say that packaged workbooks are the best & fastest way for you to get your questions answered. So doing what seems like extra work to mock up data and create a workbook that matches your situation is worth the effort and very often it's possible to mock up data and workbooks that match your situation. For example here's a question on filters using bocce ball teams where the actual use case involved hundreds of millions of records coming out of multiple data pipelines.
- Give us details of your challenge. If you can't share a workbook then we're going to need information to make up for that.
- Take screenshots. Again, a picture is worth a thousand words. For screenshots of visualizations sure that the screenshots include all of the Shelves (Rows, Columns, Filters, Pages, and the Marks Card). If you have data that is confidential on the screenshots then use an image editor to blur out those parts of the image. Please also document your data and your goal (see the Clearly document your goal section above) so we can understand what you are trying to do.
- Tell us about your data. In particular we need to know what source(s) of data you are using (Excel, Google Bigquery, ESRI Shapefiles, etc.), how you have prepared them, and roughly how many rows there are (thousands, millions, hundreds of millions, etc.).
- Describe your problem. What have you done so far, and what's not working as expected or desired?
- Clearly document your goal. See the section on that above.
PLEASE NOTE: Comments addressing the accuracy and completeness of this document are welcome here so this document can grow, and remain accurate and relevant.