[SOLR-10424] /update/docs/json is swallowing all fields

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 6.5, 7.0
• Component/s: update
• Assignee: Unassigned
• Reporter: Chris M. Hostetter
• Created: 05/Apr/17 00:21
• Updated: 10/Nov/21 19:43

Description

I'm not sure when/how exactly this broke, but sending a list of documents to /update/json/docs is currently useless – regardless of what your documents contain, all you get is 3 fields: id, _version_, and a _src_ field containing your original JSON, but none of the fields you specified are added.

Steps to reproduce...

git co releases/lucene-solr/6.5.0
...
ant clean && cd solr && ant server
...
bin/solr -e techproducts
...
curl 'http://localhost:8983/solr/techproducts/update/json/docs?commit=true' --data-binary @example/exampledocs/books.json -H 'Content-type:application/json'
...

curl 'http://localhost:8983/solr/techproducts/query?q=id:978-1933988177'
{
  "responseHeader":{
    "status":0,
    "QTime":5,
    "params":{
      "q":"id:978-1933988177"}},
  "response":{"numFound":1,"start":0,"docs":[
      {
        "id":"978-1933988177",
        "_src_":"{\n    \"id\" : \"978-1933988177\",\n    \"cat\" : [\"book\",\"paperback\"],\n    \"name\" : \"Lucene in Action, Second Edition\",\n    \"author\" : \"Michael McCandless\",\n    \"sequence_i\" : 1,\n    \"genre_s\" : \"IT\",\n    \"inStock\" : true,\n    \"price\" : 30.50,\n    \"pages_i\" : 475\n  }",
        "_version_":1563794703530328065}]
  }}

Compare with using /update/json ...

curl 'http://localhost:8983/solr/techproducts/update/json?commit=true' --data-binary @example/exampledocs/books.json -H 'Content-type:application/json'
...
curl 'http://localhost:8983/solr/techproducts/query?q=id:978-1933988177'
{
  "responseHeader":{
    "status":0,
    "QTime":0,
    "params":{
      "q":"id:978-1933988177"}},
  "response":{"numFound":1,"start":0,"docs":[
      {
        "id":"978-1933988177",
        "cat":["book",
          "paperback"],
        "name":"Lucene in Action, Second Edition",
        "author":"Michael McCandless",
        "author_s":"Michael McCandless",
        "sequence_i":1,
        "sequence_pi":1,
        "genre_s":"IT",
        "inStock":true,
        "price":30.5,
        "price_c":"30.5,USD",
        "pages_i":475,
        "pages_pi":475,
        "_version_":1563794766373584896}]
  }}

According to the ref-guide, the only diff between these two endpoints should be that /update/json/docs defaults json.command=false ... but since the top level JSON structure in books.json is a list ("[ ... ]") that shouldn't matter because that's not the solr JSON command syntax.

---

If you try to send a singular JSON document tp /update/json/docs, you get the same problem...

curl -X POST -H 'Content-type:application/json' --data-binary '{"id":"HOSS","popularity":42}' 'http://localhost:8983/solr/techproducts/update/json/docs?commit=true'
...
curl 'http://localhost:8983/solr/techproducts/query?q=id:HOSS'{
  "responseHeader":{
    "status":0,
    "QTime":0,
    "params":{
      "q":"id:HOSS"}},
  "response":{"numFound":1,"start":0,"docs":[
      {
        "id":"HOSS",
        "_src_":"{\"id\":\"HOSS\",\"popularity\":42}",
        "_version_":1563795188162232320}]
  }}

...even though the same JSON works fine to /update/json?json.command=false ...

curl -X POST -H 'Content-type:application/json' --data-binary '{"id":"HOSS","popularity":42}' 'http://localhost:8983/solr/techproducts/update/json?commit=true&json.command=false'
...
curl 'http://localhost:8983/solr/techproducts/query?q=id:HOSS'{
  "responseHeader":{
    "status":0,
    "QTime":1,
    "params":{
      "q":"id:HOSS"}},
  "response":{"numFound":1,"start":0,"docs":[
      {
        "id":"HOSS",
        "popularity":42,
        "_version_":1563795262581768192}]
  }}

Issue Links

• is related to: SOLR-11704 sample_techproducts_configs is really out-of-date
• relates to: SOLR-10581 Json update handler doesn't support nested documents in streaming mode