i have products with tags, and tags are inside tagtypes.
this is a sample document that i added to the index
{
"_index" : "products",
"_type" : "_doc",
"_id" : "1219",
"_score" : 1.0,
"_source" : {
"id" : "1219",
"product_no" : "26426492261",
"merchant_id" : 11,
"name" : "Apple »Magic Keyboard für das 12,9\" iPad Pro (4. Generation)« iPad-Tastatur",
"category" : "Technik>Multimedia>Zubehör>Tastatur>iPad Tastatur",
"deep_link" : "https://foo",
"short_description" : null,
"long_description" : "Apple:",
"brand" : "Apple",
"merchant_image_url" : "http://something",
"tagtypes" : [
[
{
"Memory" : [ ]
}
]
]
}
},
That tagtype "Memory" is dynamically created while indexing the products.
I tried to add tags to that key
//attach tags also to ES
$params = [
'index' => 'products',
'id' => $product['_id'],
'body' => [
'script' => [
'source' => 'if (!ctx._source.tagtypes.'.$tagType->name.'.contains(params.tag)) { ctx._source.tagtypes.'.$tagType->name.'.add(params.tag) }',
'lang' => 'painless',
'params' => [
'tag' => $tag->value
]
]
]
];
But i receive an error like
{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"failed to execute script"}],"type":"illegal_argument_exception","reason":"failed to execute script","caused_by":{"type":"script_exception","reason":"runtime error","script_stack":["if (!ctx._source.tagtypes[\"Memory\"].contains(params.tag)) { "," ^---- HERE"],"script":"if (!ctx._source.tagtypes[\"Memory\"].contains(params.tag)) { ctx._source.tagtypes[\"Memory\"].add(params.tag) }","lang":"painless","position":{"offset":16,"start":0,"end":60},"caused_by":{"type":"wrong_method_type_exception","reason":"cannot convert MethodHandle(List,int)int to (Object,String)String"}}},"status":400}
Could anyone help me with that. I couldnt find any documentation about it, as the examples are often too basic.
Is it generally possible to save to "deeper keys" like this ? Or can i just create "tags" as simple list (without any depth)
Thanks in advance Adrian!
Your field tagtypes
is an array of arrays of objects which themselves contain one-key arrays.
When you're dealing with such "deep" structures, you'll need some form of iteration to update them.
For
loops are a good place start but they often lead to java.util.ConcurrentModificationExceptions
. So it's easier to work with temporary copies of data and then replace the corresponding _source
attribute when done with the iterations:
{
"query": {
"match_all": {}
},
"script": {
"source": """
if (ctx._source.tagtypes == null) { return; }
def originalTagtypes = ctx._source.tagtypes;
def newTagtypes = [];
for (outerGroup in originalTagtypes) {
// keep what we've got
newTagtypes.addAll(outerGroup);
// group already present?
def atLeastOneGroupContainsTag = outerGroup.stream().anyMatch(tagGroup -> tagGroup.containsKey(params.tag));
// if not, add it as a hashmap of one single empty list
if (!atLeastOneGroupContainsTag) {
Map m = new HashMap();
m.put(params.tag, []);
newTagtypes.add(m);
}
}
ctx._source.tagtypes = [newTagtypes];
""",
"lang": "painless",
"params": {
"tag": "CPU"
}
}
}
which'll end up updating the tagtypes
like so:
{
...
"tagtypes" : [
[
{
"Memory" : [ ]
},
{
"CPU" : [ ] <---
}
]
],
...
}
You're right when you say that the documentation examples are too basic. Shameless plug: I recently published a handbook that aims to address exactly that. You'll find lots non-trivial scripting examples to gain a better understanding of the Painless scripting language.