Create a Model from a File

You can create a model using personal files as a data source.

Context

The data is analyzed first so that the application can suggest dimensions for the model you want to create. You can then refine the suggestions by specifying dimension types and fixing any data-quality problems.

The overall workflow is as follows:
  1. Import a source file.
  2. Choose dimension types for the dimensions in your new model.
  3. Specify attributes for the dimensions.
  4. Cleanse your data and fix any mapping or data quality problems.

Data files can be in your local file system or in your network. The source data can be an Excel spreadsheet (.xlsx) or a delimited text file (.csv or .txt). If you import data from Microsoft Excel, and if the data is saved in separate sheets in the Excel workbook, either you can choose which sheet to import (if from a local file system) or the first sheet is automatically imported (from a network file).

The source data must include columns that can be used to create dimensions in a new model, and it must also include transactional data (measures). The transactional data is visible in the Data Foundation view, which shows you the fact table containing the raw, unaggregated transactional data loaded into your model. For details, see Learn About the Modeler.

The data will typically include details for the main account. You can identify which dimension is the main account during the import process, but if no account data is found, an account dimension will be generated automatically.

In addition to the main dimensions and transactional data, the source can also include attributes for dimensions; that is, text information such as hierarchy information or other properties.

Procedure

  1. From the Modeler start page, select From a CSV or Excel File.
  2. In the Import Data From File dialog, choose whether you want to import data from a file on your local system, or from a file server.

    If you don't see the options to choose a local system or file server, see Allow Data Import and Model Export with a File Server.

    Tip
    If you import a file from a file server, you can also schedule imports from that file. For more information, see Update and Schedule Models.
  3. If you're importing from a file server, choose a file server connection, or select Create New Connection.

    If you create a new file server connection, specify the path to the folder where the data files are located. For example: C:\folder1\folder2 or \\servername\volume\path or /mnt/abc.

  4. Select the file you want to import.
  5. Optional: If you are importing from a local Excel workbook containing multiple sheets, select the sheet you want to import.

    If you are importing from an Excel file on a file server, the first sheet is automatically imported.

  6. Optional: If you're importing a .csv file, select which delimiter is used in the file, or select Auto-detect.
  7. Click Import.
    For small data files, the data is imported and displayed in the data integration view, which shows the imported data columns that you'll define as dimensions, measures, and attributes.

    Larger data files may take some time to upload. You can work on other tasks while the data is being uploaded in the background.

  8. Once the data upload is complete, select it from the Draft Data list to open it in the data integration view.

Next Steps

Video: How to Create Models from Files

Open this video in a new window

In this video, you will create a model based on a file, review data wrangling and transformation options available, validate the data, and review the options available to edit and enhance saved models.

After the initial import of raw data, continue with the data preparation task before completing your model. See Import and Prepare Fact Data for a Classic Account Model, and Import and Prepare Fact Data for a Model with Measures.
Related Information