Share via

Unpack Zipped Datasets

  • Article

Unpacks datasets from a zip package in user storage

Category: Data Input and Output

Note

Applies to: Machine Learning Studio (classic) only

Similar drag-and-drop modules are available in Azure Machine Learning designer.

Module overview

This article describes how to use the Unpack Zipped Datasets module in Machine Learning Studio (classic), to upload data and script files in compressed format, and then unzip them for use in an experiment.

The purpose of this module is to reduce data transfer times when working with very large datasets by saving and uploading your data files in a compressed format. Generally, zipping files is a good option when your dataset is so large that you want to use compression for the upload, to minimize upload time and associated costs.

The module takes as input a dataset in your workspace. The dataset must have been uploaded in a compressed format. The module then decompresses the dataset and adds the data to your workspace.

How to use Unpack Zipped Datasets

This section describes how to prepare your data and then unzip it in Machine Learning Studio (classic).

Step 1. Prepare files

Before uploading your file, make sure that the data in the file can be used in Machine Learning:

  • Ensure that the data in the file uses UTF-8 encoding.

    If the file is small enough, you can open it in Notepad and then save the file in the desired encoding. Many other text editors offer similar functionality. For CSV files, you can use Excel's Save As or Export commands to specify a file format and encoding.

  • Verify that the data files use a supported format, such as CSV, TSV, ARFF, or SVMLight.

  • Compress the data by adding the data file to a .ZIP or .GZ format archive file. Other archive types are not supported.

  • Remove password protection. If any of the files or the compressed folder itself has been encrypted or password-protected, you must unlock or decrypt the file before you upload it. The module cannot detect encrypted data types and does not support dialog boxes for password entry from arbitrary clients.

Step 2. Upload dataset to your workspace

Next, upload the zipped dataset to your experiment workspace.

  1. Click NEW, select DATASET, and select FROM LOCAL FILE.

  2. Locate the zipped file to upload. When you select the file, the type should automatically be set to Zip file (.zip).

Step 3. Add zipped dataset to experiment

After the dataset has uploaded completely, add it to your experiment in zipped format.

  1. In the left-hand navigation pane of Machine Learning Studio (classic), select Saved Datasets, and then expand My Datasets.

  2. Locate the zipped dataset that you just uploaded, and drag it to the experiment canvas.

Step 4. Unpack dataset

The final step is to unpack the dataset.

  1. Connect the zipped dataset to the input of the Unpack Zipped Datasets module.

  2. In Dataset to Unpack, type the name of a single dataset to unpack.

    • If you saved a worksheet with the name Sheet1 as an Excel CSV file named Test.csv, the name of the dataset would be Test.csv, not Sheet1.

    • The name that you type in the Dataset to Unpack text box must be exactly the same as the name of the original file before it was compressed, including the file name extension. For example, if you want to unpack a dataset based on the text file Users.txt, type Users.txt, not Users.

    • If you put multiple files into one compressed folder, you must unpack one dataset at a time.

    Tip

    If you leave the property blank, the module gets the file name from the zipped file, assuming the compressed archive file contains only one source file. If the compressed archive contains multiple files, a run-time error is raised.

  3. For Dataset file format, specify the original format of the dataset: that is, the format before it was zipped.

    You can upload and unzip datasets that were created using any of these formats: CSV, ARFF, TSV, SvmLight.

    If this property is left empty, the module identifies the dataset using the source file name.

  4. Select the option, File has header row, if the original dataset had a header row. Otherwise the first row of data is used as the header. If this is not what you want, add a header prior to input.

    This option applies only to .CSV and .TSV files.

    Note

    If you change the format of the file, this option is reset.

  5. If the file is compressed, use the Compression file format option to specify the algorithm that was used to compress or expand the file.

    Currently the .ZIP and GZ (or Gzip) formats are supported.

  6. Run the experiment.

Results

  • To verify that the data was imported correctly, right-click the Unpacked Zipped Datasets module, and select Visualize .

  • To change the name of the dataset, right-click the Unpacked Zipped Datasets module, and select Save as Dataset. At this point you can type a different name.

    This option is handy if you are unpacking multiple datasets from a single ZIP file.

Examples

To demonstrate how this module works, we created a sample .ZIP file containing four different CSV files. All files were saved from Excel.

File name Description
names-uni.csv Unicode file with column headings
names-utf.csv UTF-8 file with column headings
nonames-uni.csv Unicode file with no column headings
nonames-utf8.csv UTF-8 file with no column headings

The entire zipped file was uploaded, and then the Unpack Zipped Datasets module was run four times to extract each of the four files, using these settings:

  1. Dataset to unpack = names-uni.csv, File has header row = TRUE
  2. Dataset to unpack = names-utf8.csv, File has header row = TRUE
  3. Dataset to unpack = nonames-uni.csv, File has header row = FALSE
  4. Dataset to unpack = nonames-utf8.csv, File has header row = FALSE

The results were as expected:

File name Upload result
names-uni.csv Error 0049: Error while parsing the file. File is not Unicode (UTF-8) encoded
names-utf8.csv Success. Uses original column names from source file.
nonames-uni.csv Error 0049: Error while parsing the file. File is not Unicode (UTF-8) encoded
nonames-utf8.csv Success. Column names Col1, col2, ...coln are automatically added to the dataset.

Note

If you use the option, File has header row = TRUE, and the source file actually does not have a column heading, the first row of data is used as the column heading.

Technical notes

You cannot use this module to unpack zipped R packages into your workspace. R packages must be uploaded and consumed as zipped files.

For more information about how to work with zipped R packages, see Execute R Script.

Note

Confused about the difference between UTF-8 and Unicode? See this Wikipedia article: What is UTF-8

Module parameters

Name Range Type Default Description
Compression file format Zip

Gzip		 compression rule Zip Compression algorithm used to compress or expand the file.
Dataset to Unpack Any String none Name of dataset to register with Azure ML Studio (classic). If the name of a dataset is not specified, the name is obtained from the file name in the zipped file.
Dataset file format CSV

TSV

ARFF

SVMLIGHT		 File format CSV File format of the dataset in the zipped file
File has header row TRUE/FALSE Boolean False Set to True only if the CSV/TSV file has a header row

Expected inputs

Name Type Description
Dataset Zip Zipped file containing datasets

Output

Name Type Description
Results dataset Data Table Output dataset

See also

Data Input and Output