Advanced Querying

Most OrderCloud resources have list endpoints that can return thousands, even hundreds of thousands of items. We understand that effectively querying this data is not only practical but necessary and have built an assortment of API tools to help pinpoint and extract the exact data you and your users need. Up first, let's talk about Search.

Searching

Google-esque searches, simply provide a search term.

This type of querying is best for natural language searches like a product search on a buyer site. Our search handles case-insensitive word stemming (plural inference for example), synonyms, and intelligently ranks results based on fields relevant to each resource. If you're curious about which default fields are taken into account and how much weight they're given for a search check out our API reference.

1GET https://api.ordercloud.io/v1/me/products?search=bike HTTP/1.1

This will return all results such as bike, words including bike, bikes, bicycle, etc. The fields searched include ID,Name, and Description. (This of course varies from endpoint to endpoint). Fields marked "Searchable" in the API reference will be searched and weighted according to priority level.

You can also optionally specify which fields (of the searchable fields available) are searched by providing a searchOn parameter for any field available on the resource you're searching including extended properties (XP). Here's we're searching on Name and xp.Color so if our search finds a match in either of those two fields we will get a result.

1GET https://api.ordercloud.io/v1/me/products?search=bike&searchOn=Name,xp.Color HTTP/1.1

Filtering

Filtering is best for looking at specific values in a given field. If you do not require natural language type searching over text fields, filtering is a more performant and more powerful method for querying data.

Filtering On XP

Let's examine a common scenario of filtering on an extended property (XP) field. Suppose our XP field looks like this:

1{
2  "xp": {
3    "MoreInfo": {
4      "TeamName": "Tigers",
5      "Gender": "Male"
6    }
7  }
8}

We can use dot notation to access deeply nested values. The following call will return all buyers with team name "Tigers"

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?xp.MoreInfo.TeamName=Tigers HTTP/1.1

Fuzzy Searches

Fuzzy matches are supported using the * wildcard character.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?LastName=*Smith&FirstName=John* HTTP/1.1

This will return both "John Smith" and "Johnny McSmooth".

Logical OR

You can also use | as a logical OR.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?LastName=Smith|Jo*&FirstName=Johnny HTTP/1.1

This will return "John Smith", "Johnny Jones", and "John Johnson". Maybe you want "John Smith" but not "John Jones".

Negate and Logical AND

You can negate your conditions by prefixing them with !, and logically AND them together by simply providing the same parameter multiple times.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?LastName=!Smith&LastName=!Jones HTTP/1.1

This will return all users except those with last name "Smith" or "Jones".

Comparison Operators

Dates and numeric values support > (greater than) and < (less than) prefixes.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?DateCreated=>2015-01-01 HTTP/1.1

For a more advanced example, let's say you want users whose ID is the range of 0 to 9 inclusive. Ranges are not directly supported, but you can use the existing features to achieve this.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?ID=0|1|2|3|4|5|6|7|8|9 HTTP/1.1

Of course that's not going to be feasible if the range is very large or you're dealing with floating-point numbers. A better way would be to leverage > and <, but we want the range to be inclusive, and there are no >= or <= operators. We can however leverage the ! operator.

Here we're saying "give me all users whose ID is not less than 0 and not greater than 9", which is effectively equivalent to our 0-9 range.

1GET https://api.ordercloud.io/v1/buyers/{buyerID}/users?ID=!<0&ID=!>9 HTTP/1.1

Absence of a value

Sometimes you need a list of items where some field either doesn't exist or is null or undefined. To search on the absence of a field, you can combine the negate operator and the wildcard operator.

1GET https://api.ordercloud.io/v1/orders?xp.Billed=!* HTTP/1.1

Here we're saying "give me all orders that haven't been billed (xp.Billed is undefined, null, or doesn't exist)"

Note: Searching on the absence of a value can be fairly expensive since we're using the wildcard filter *. Another (more performant) way to solve the same problem would be to initialize your field with some value that you define as being "null" and then simply filter on that value. Prefer this method when possible.

Performance Considerations

Your data is highly indexed for fast retrieval using the methods mentioned above, including deep XP object graphs. However, there are a few important takeaways to keep search and listing functionality efficient. Filters with many OR conditions or with values that start with the * wildcard character may be particularly performance-sensitive. The larger the set of data, the more likely you are to notice any performance impact. It pays to know the data model well, and use grouping constructs like User Groups and Categories effectively.

Sorting

In addition to being able to narrow down the data itself, you can also indicate to OrderCloud how you want it sorted. Perhaps you want to sort the results (in ascending order) by the last name of the person who placed an order. To do this you can simply include the sortBy parameter for the field you would like to sort by.

1GET https://api.ordercloud.io/v1/me/orders/incoming?sortBy=LastName HTTP/1.1

If you want to sort results in descending order simply add the ! operator

1GET https://api.ordercloud.io/v1/me/orders/incoming?sortBy=!LastName HTTP/1.1

Combining Queries

Searching, filtering and sorting can all be mixed and matched to give you ultimate control when defining what is in a list and how that list is presented to your users.

Pagination

All list responses will return an object with two values: Meta and Items. The Meta value returns important information regarding your query, such as total results, how many results are shown in the Items array, how many more pages of results there are, and what page you are currently on. The Items array stores the actual results of the resource you are querying.

1HTTP/1.1 200 OK
2Content-Type: application/json; charset=UTF-8
3
4  {
5    "Meta": {
6      "Page": 1,
7      "PageSize": 20,
8      "TotalCount": 25,
9      "TotalPages": 2,
10      "ItemRange": [1, 20]
11    },
12    "Items": ["..."]
13  }

Lists default to 20 items per page, but you can set the pageSize up to 100. If you have more results than the page count, you will have multiple pages. You can use the Page and PageSize values from the Meta object to specify the page (starting at 1) and items per page.

Note: Unlike the page parameter, ItemRange values are zero-based.

LastID method

When it's necessary to paginate through many items (generally anything > page 30) we highly recommend using the following pattern for the optimal performance:

1. Make your first list call with page=1&pageSize=100&sortBy=ID as query parameters

2. Keep track of the ID of the last item returned in the list

3. In subsequent calls, always request page=1&pageSize=100&sortBy=ID&ID=>{lastID}

4. Stop paginating when the query returns an empty Items array

Important note: For resources using premium search (products, and soon orders), this method yields unexpected results if you store IDs as different number types. For example 82020 and 85919200509 cannot be accurately compared and sorted because the first ID is identified as an integer and the second as a long. To work around this limitation you could store IDs with leading 0s so they are all identified as the same type. So instead of 82020, you would set the ID to 00000082020.

Still have questions?
Ask in our Community Channel

Content Powered By