tableschema-spss 0.2.0

No project description provided

Read and write between SPSS and Table Schema.

Getting Started

Installation

pip install tableschema-spss

Storage

Package implements the Tabular Storage interface.

With a base path

We can get storage with a specified base path this way:

from tableschema_spss import Storage

storage_base_path = 'path/to/storage/dir'
storage = Storage(storage_base_path)

We can then interact with storage buckets (‘buckets’ are SPSS .sav/.zsav files in this context):

storage.buckets  # list buckets in storage
storage.create('bucket', descriptor)
storage.delete('bucket')  # deletes named bucket
storage.delete()  # deletes all buckets in storage
storage.describe('bucket') # return tableschema descriptor
storage.iter('bucket') # yields rows
storage.read('bucket') # return rows
storage.write('bucket', rows)

Without a base path

We can also create storage without a base path this way:

from tableschema_spss import Storage

storage = Storage()  # no base path argument

Then we can specify SPSS files directly by passing their file path (instead of bucket names):

storage.create('data/my-bucket.sav', descriptor)
storage.delete('data/my-bucket.sav')  # deletes named file
storage.describe('data/my-bucket.sav') # return tableschema descriptor
storage.iter('data/my-bucket.sav') # yields rows
storage.read('data/my-bucket.sav') # return rows
storage.write('data/my-bucket.sav', rows)

Note that storage without base paths does not maintain an internal list of buckets, so calling storage.buckets will return None.

Reading .sav files

When reading SPSS data, SPSS date formats, DATE, JDATE, EDATE, SDATE, ADATE, DATETIME, and TIME are transformed into Python date, datetime, and time objects, where appropriate.

Other SPSS date formats, WKDAY, MONTH, MOYR, WKYR, QYR, and DTIME are not supported for native transformation and will be returned as strings.

Creating .sav files

When creating SPSS files from Table Schemas, date, datetime, and time field types must have a format property defined with the following patterns:

• date: %Y-%m-%d

• datetime: %Y-%m-%d %H:%M:%S

• time: %H:%M:%S.%f

Table Schema descriptors passed to storage.create() should include a custom spss:format property, defining the SPSS type format the data is expected to represent. E.g.:

{
    "fields": [
        {
            "name": "person_id",
            "type": "integer",
            "spss:format": "F8"
        },
        {
            "name": "name",
            "type": "string",
            "spss:format": "A10"
        },
        {
            "type": "number",
            "name": "salary",
            "title": "Current Salary",
            "spss:format": "DOLLAR8"
        },
        {
           "type": "date",
           "name": "bdate",
           "title": "Date of Birth",
           "format": "%Y-%m-%d",
           "spss:format": "ADATE10"
        }
    ]
}