Leave a comment at the end of this page or email contact@krishagni.com
Query
- Vinayak Pawar
- Swapnil Ingle (Unlicensed)
Executing Any AQL
Below API can be used to execute any AQL and obtain the desired results. AQL is custom query language built-in OpenSpecimen to allow power users perform complex queries on data without requiring to know anything about the OpenSpecimen data model.
| Item | Value |
|---|---|
| HTTP Method | POST |
| URI | /rest/ng/query |
| Body | {
"cpId": <cpId>,
"aql": <AQL>,
"wideRowMode": <OFF|SHALLOW|DEEP>,
"outputColumnExprs": <true|false>,
"outputIsoDateTime": <true|false>,
"timeoutInSeconds": <integer>
}
|
| Response | JSON containing desired results |
Parameters
| Parameter | Description |
|---|---|
| cpId | Collection protocol ID. If specified, limits the results to specified CP records. Useful when AQL contains CP specific custom fields. |
| aql | Valid AQL string that needs to be executed |
| wideRowMode | Specifies whether multi-valued attributes result in a single row or one row per value. Default value is OFF. Other permitted values are SHALLOW and DEEP. Try out to see what fits best for your use case. |
| outputColumnExprs | Specifies whether the column labels or AQL expression needs to be included in the query response. By default, user friendly column labels are included in query response. |
| outputIsoDateTime | Specifies how the date column values needs to be serialised in the query response. If true, then date/time values are serialised using ISO format: yyyy-MM-dd'T'HH:mm:ss. Otherwise, date/time values are serialised using the format specified in OS locale settings |
timeoutInSeconds | Specifies how long to wait for the query to complete before aborting it. When not specified, this field defaults to 55 seconds. This is a guard against the run away query that hog the system resources, eventually resulting in service disruption. To wait forever for the query to complete, use -1. This should be avoided unless you are aware about the consequences of allowing a runaway query. |
Example
Given below is an example query request to obtain count of aliquots by visit dates:
{
"cpId": 1,
"aql": "select Participant.ppid, SpecimenCollectionGroup.collectionDate, count(distinct Specimen.id) where Specimen.lineage = \"Aliquot\"",
"outputIsoDateTime": true
}
Sample response
Getting List of all saved-queries
Request Details
| Item | Value |
|---|---|
| HTTP Method | GET |
| URI | /rest/ng/saved-queries |
| Response | JSON containing detailed list of saved queries |
Parameters
| Parameter | Description |
|---|---|
| cpId | The cpId is used to filter the saved queries based on the CP. This is helpful in running the queries in background using scheduled jobs for initialisation of CP specific custom fields/forms. Note: This is the CP that you select in the top-left dropdown of the add/edit query wizard. |
searchString | Matches a substring in the title of the saved query. |
start | Used for paginating the results. If start=5, the output result will start from row number 5. |
max | Used for paginating the results. If max=10, the output result will be maximum 10 rows |
countReq | Boolean type parameter, if specified as 'true' output will show the total number of saved-queries. Otherwise it won't. |
Sample response
Get URL: 'rest/ng/saved-queries?countReq=true&start=0&max=10'
Executing Saved Query
Request Details
| Item | Value |
|---|---|
| HTTP Method | POST |
| URI | /rest/ng/query/{savedQueryId} |
| Body | {
"drivingForm":"Participant",
"wideRowMode": "DEEP",
"startAt": 0,
"maxResults": 20
}
|
| Response | JSON containing the output of the saved queries |
Parameters
Note: The savedQueryId placeholder should be replaced with the actual ID of the saved query.
| Parameter | Description | Allowed Values |
|---|---|---|
drivingForm | Driving form determines the search perspective. When left empty, it defaults to Participant. (For example when drivingForm is Participant, the root table is 'catissue_coll_prot_reg' which is then used to join with the other tables.Similarly when drivingForm is Specimen, the root table will be catissue_specimen, which will be joined with the other tables.) | Participant, Specimen |
startAt | Used for paginating the results. If startAt=5, the output result will startAt from row number 5. | Integer |
maxResults | Used for paginating the results. If maxResults=10, the output result will be maximum 10 rows | Integer |
wideRowMode | Specifies whether multi-valued attributes result in a single row or one row per value. Default value is OFF. Other permitted values are SHALLOW and DEEP. Try out to see what fits best for your use case. | DEEP, SHALLOW, OFF |
Sample response
Exporting Query Data
Request
| Method | POST |
| URL | rest/ng/query/export |
| Body | {
"aql": "select Participant.ppid, Participant.regDate, Participant.gender, Participant.race, Specimen.label, Specimen.type where Participant.participantId exists",
"wideRowMode": "DEEP"
}
|
| Response | {
"dataFile": "query_0_61e59bc2-9350-47db-9af8-a1e7a3b65bc2",
"completed": false
}
|
Parameters
| aql | Valid AQL query whose output needs to be exported to a CSV file |
| dataFile | Handle or file ID to use for downloading the exported CSV file |
| completed | Boolean indicating whether the exported CSV file can be downloaded right away |
Download Exported Data File
Request
| Method | GET |
| URL | rest/ng/query/export?fileId=<dataFile> |
| Response | Scenario 1: When the exported data file is not ready HTTP 400 with the following error message in the response body [
{
"code":"QUERY_EXPORT_DATA_IN_PROGRESS",
"message":"The query result data export is in progress. Retry downloading the file after some time."
}
]
Scenario 2: When the exported data file is ready HTTP 200 with the ZIP binary stream in the response body |
Parameters
| fileId | The handle or file ID returned by the export API |
Query Export in Action
Initiate Export
Download Exported File - Attempt 1 of N
Download Exported File
FAQ
Let’s assume we've the following data:
A specimen record with label L
Specimen L has 2 biohazards - H1, H2
Specimen L has 2 frozen events F1, F2
Output includes specimen label (L), biohazards, and one or more fields of frozen event.
- What does OFF, SHALLOW and DEEP do? What are the differences between them?
Let's assume we've a specimen record L with biohazards H1 and H2. Also, let's assume the specimen L has 2 frozen events F1 and F2.Using above, it is easy to explain the various wide row modes:
- wideRowMode = OFF. There will be one row for each combination of biohazard and frozen event. That is -
(L, H1, F1), (L, H1, F2), (L, H2, F1), (L, H2, F2)
- wideRowMode = SHALLOW. There will be one row for each frozen event. That is -
(L, H1, H2, F1), (L, H1, H2, F2)
- wideRowMode = DEEP. There will be one row for each specimen irrespective of the number of biohazards and frozen events. That is -
(L, H1, H2, F1, F2)
- wideRowMode = OFF. There will be one row for each combination of biohazard and frozen event. That is -
Why is it so hard to get the same results from the same query in both the API and in OpenSpecimen?
The UI uses either SHALLOW or DEEP. It does not use OFF.
When the “Enable Wide Rows” checkbox is ticked, the wideRowMode=DEEP. When it is unticked, wideRowMode=SHALLOW.
Leave a comment at the end of this page or email contact@krishagni.com