SODA API - Create new dataset with columns?

[warmfusion]

Anyone found how to create new datasets from the API such that I don't have to manually create columns and configuration properties?

[ldodds]

I think this is possible but it as not documented anywhere. The Java client
library seems to have features, including editing dataset metadata that
isn't in the other libraries.

See code/comments here:

https://github.com/socrata/soda-java

I've not reverse engineered the protocol yet though and can't find any docs

L.
On 26 Aug 2014 21:58, "Toby Jackson" [email protected] wrote:

Anyone found how to create new datasets from the API such that I don't
have to manually create columns and configuration properties?

—
Reply to this email directly or view it on GitHub
#1.

[warmfusion]

Interesting, I've been scraping the AJAX calls from the website and found you might be able to GET/POST/PUT on https://data.bathhacked.org/views//columns.json

For example, when I added a test column, the following was POSTed to columns.json

{"name":"test","description":"Description","dataTypeName":"text","format":{},"metadata":{},"width":100}

[warmfusion]

Creating a new view through the website

POST https://data.bathhacked.org/views.json

Data:

{"name":"Dataset-Title","description":"A Brief Description","metadata":{"renderTypeConfig":{"visible":{"table":true}},"custom_fields":{"Author":{"Author Name":""},"Publication":{"Publication Date":"","Update Frequency":""},"Additional Licence Detail":{"Additional Licence Information":"","Licence URL":"","Additional Attribution Statement":""}},"availableDisplayTypes":["table","fatrow","page"],"jsonQuery":{}},"flags":["dataPublicRead"],"displayType":"table","displayFormat":{},"query":{}}

Response

{
  "id" : "72vp-je3j",
  "name" : "Dataset-Title",
  "averageRating" : 0,
  "createdAt" : 1409089331,
  "description" : "A Brief Description",
  "displayType" : "table",
  "domainCName" : "data.bathhacked.org",
  "newBackend" : false,
  "numberOfComments" : 0,
  "oid" : 14850,
  "publicationAppendEnabled" : false,
  "publicationGroup" : 10985,
  "publicationStage" : "unpublished",
  "tableId" : 10985,
  "totalTimesRated" : 0,
  "viewLastModified" : 1409089331,
  "viewType" : "tabular",
  "columns" : [ ],
  "grants" : [ {
    "inherited" : false,
    "type" : "viewer",
    "flags" : [ "public" ]
  } ],
  "metadata" : {
    "custom_fields" : {
      "Additional Licence Detail" : {
        "Additional Attribution Statement" : "",
        "Licence URL" : "",
        "Additional Licence Information" : ""
      },
      "Publication" : {
        "Publication Date" : "",
        "Update Frequency" : ""
      },
      "Author" : {
        "Author Name" : ""
      }
    },
    "renderTypeConfig" : {
      "visible" : {
        "table" : true
      }
    },
    "availableDisplayTypes" : [ "table", "fatrow", "page" ]
  },
  "owner" : {
    "id" : "tnwz-gh9e",
    "displayName" : "Toby Jackson",
    "emailUnsubscribed" : false,
    "roleName" : "editor",
    "screenName" : "Toby Jackson",
    "rights" : [ "create_datasets", "view_domain", "create_pages", "view_goals", "view_dashboards" ]
  },
  "query" : {
  },
  "rights" : [ "read", "write", "add", "delete", "grant", "add_column", "remove_column", "update_column", "update_view", "delete_view" ],
  "tableAuthor" : {
    "id" : "tnwz-gh9e",
    "displayName" : "Toby Jackson",
    "emailUnsubscribed" : false,
    "roleName" : "editor",
    "screenName" : "Toby Jackson",
    "rights" : [ "create_datasets", "view_domain", "create_pages", "view_goals", "view_dashboards" ]
  },
  "flags" : [ "default" ]
}

[markowen]

Good work on this guys. I'd assumed that there was no mechanism for manipulating datasets. I'll have a dive into this & see if we can find something interesting.

[ldodds]

It looks like we may also be able to automatically snapshot, create working copies and publish datasets through the API too, which might be useful for bulk updates.

[warmfusion]

I started work on some ruby SODAClient extensions last night to try and expand the available methods to incorporate some of these newly scraped behaviours.

My target is to effectively solve my open issue on socrata/soda-ruby#13