Standardize Data with Mapping Tables

Picture this: you are in the Report Builder, building a Revenue by State report. You are in the zone. Everything is going swimmingly until you go to add a billing state grouping to your report and you see this:

How could this happen?

Unfortunately, a lack of standardization can sometimes lead to messy data and headaches when building reports. In our example, there may not have been a dropdown menu or standardized way for our customers to input their billing state information. This lead to a variety of values - pa, PA, penna, pennsylvania, and Pennsylvania - all for the same state, which will certainly lead to some strange results in the Report Builder.

It is possible that there is a tech resource that can help you clean up the data or insert the columns you need directly into your database, but if not, we have another solution: the mapping table. A mapping table allows you to quickly and easily cleanse and standardize any messy data by mapping data to a single output.

How do I create it?

Data formatting refresher:

• Make sure your spreadsheet has a header row.
• Avoid using commas! It will cause problems when you upload the file.
• Use the standard date format (YYYY-MM-DD HH:MM:SS) for dates.
• Percentages must be entered as decimals.
• Make sure any leading or trailing zeroes are properly retained.

Before you dive in, we recommend that you export the raw table data. Looking at the raw data first means you can explore all possible combinations for the data you need to clean up, thereby ensuring that the mapping table covers everything.

To make a mapping table, you need to create a two column spreadsheet that follows the formatting rules for file uploads.

In the first column, enter the values stored in your database with only one value in each row. For example, pa and PA cannot be on the same line - each input needs to have its own row. See below for an example.

In the second column, enter what these values should be. Continuing with our billing state example, if we want pa, PA, Pennsylvania, and pennsylvania to simply be PA, we would enter PA in this column for each input value.

What do I need to do in MBI to use it?

After you have finished creating the mapping table, you’ll need to upload the file into MBI and create a joined column that relocates the new field into the desired table. You can do this after the file is synced to your data warehouse.

In our example, we will move the column we created on the mapping_state table (state_input) to the customer_address table using a joined column. This will allow us to group by the clean state_input column in our reports instead of the state column.

To create the joined column, navigate to the table that the field will be relocated to in the Data Warehouse Manager. In our example, this would be the customer_address table.

1. Click Create a Column.

2. Select Joined Column from the Definition dropdown.

3. Give the column a name that differentiates it from the state column in your database. We will go with billing state (mapped) so we can tell which column to use when segmenting in the report builder.

4. The path we need to connect the tables does not exist, so we need to create a new one. Click Create new path in the Select a table and column dropdown.

   If you are not sure what the table relationship is or how to properly define the primary and foreign keys, check out our tutorial for some help.

   • On the Many side, select the table you are relocating the field to (again, for us it is customer_address) and the Foreign Key column, or state column, in our example.

   • On the One side, select the mapping table and the Primary key column. In this case, we would select the state_input column from the mapping_state table.

   • Here is a look at what our path looks like:

     

5. When finished, Click Save to create the path.

6. The path may not populate immediately after saving - if this happens, click the Path box and select the path you just created.

7. Click Save to create the column.

That is it!

What do I do now?

After an update cycle completes, you will be able to use your new joined column to properly segment your data instead of the messy column from your database. Take a look at our grouping options now - no more stress mess:

Mapping tables are handy for any time you want to clean up some potentially messy data in your data warehouse. However, mapping tables can also be used for some other cool use cases, like replicating your Google Analytics channels in MBI.

Related

• Understanding and evaluating table relationships
• Creating/deleting paths for calculated columns

Business.Adobe.com resources