fix(parsers.xpath): Improve handling of complex-type nodes (#13660)

go.mod

@@ -27,7 +27,7 @@ require (
 	github.com/alecthomas/units v0.0.0-20211218093645-b94a6e3cc137
 	github.com/aliyun/alibaba-cloud-sdk-go v1.62.470
 	github.com/amir/raidman v0.0.0-20170415203553-1ccc43bfb9c9
-	github.com/antchfx/jsonquery v1.3.2
+	github.com/antchfx/jsonquery v1.3.3
 	github.com/antchfx/xmlquery v1.3.17
 	github.com/antchfx/xpath v1.2.4
 	github.com/apache/arrow/go/v13 v13.0.0-20230630125530-5a06b2ec2a8e

go.sum

@@ -182,8 +182,8 @@ github.com/andybalholm/brotli v1.0.0/go.mod h1:loMXtMfwqflxFJPmdbJO0a3KNoPuLBgiu
 github.com/andybalholm/brotli v1.0.1/go.mod h1:loMXtMfwqflxFJPmdbJO0a3KNoPuLBgiu3qAvBg8x/Y=
 github.com/andybalholm/brotli v1.0.5 h1:8uQZIdzKmjc/iuPu7O2ioW48L81FgatrcpfFmiq/cCs=
 github.com/andybalholm/brotli v1.0.5/go.mod h1:fO7iG3H7G2nSZ7m0zPUDn85XEX2GTukHGRSepvi9Eig=
-github.com/antchfx/jsonquery v1.3.2 h1:/BgHv1le9CCkqDe7t1x5BRlCg6DQmXTsztnMQFG5Hoc=
+github.com/antchfx/jsonquery v1.3.3 h1:zjZpbnZhYng3uOAbIfdNq81A9mMEeuDJeYIpeKpZ4es=
-github.com/antchfx/jsonquery v1.3.2/go.mod h1:VsW9O/sNgHoUVvhoMEjR+opjIOjKOViNFTpAlxcI4Ws=
+github.com/antchfx/jsonquery v1.3.3/go.mod h1:1JG4DqRlRCHgVYDPY1ioYFAGSXGfWHzNgrbiGQHsWck=
 github.com/antchfx/xmlquery v1.3.17 h1:d0qWjPp/D+vtRw7ivCwT5ApH/3CkQU8JOeo3245PpTk=
 github.com/antchfx/xmlquery v1.3.17/go.mod h1:Afkq4JIeXut75taLSuI31ISJ/zeq+3jG7TunF7noreA=
 github.com/antchfx/xpath v1.1.7/go.mod h1:Yee4kTMuNiPYJ7nSNorELQMr1J33uOpXDMByNYhvtNk=

plugins/parsers/xpath/parser.go

@@ -4,6 +4,7 @@ import (
 	"encoding/hex"
 	"errors"
 	"fmt"
+	"reflect"
 	"strconv"
 	"strings"
 	"time"
@@ -448,11 +449,21 @@ func (p *Parser) parseQuery(starttime time.Time, doc, selected dataNode, config
 					}
 				}
 
-				if config.FieldsHexFilter != nil && config.FieldsHexFilter.Match(name) {
+				// Handle complex types which would be dropped otherwise for
-					if b, ok := v.([]byte); ok {
+				// native type handling
-						v = hex.EncodeToString(b)
+				if v != nil {
+					switch reflect.TypeOf(v).Kind() {
+					case reflect.Array, reflect.Slice, reflect.Map:
+						if b, ok := v.([]byte); ok {
+							if config.FieldsHexFilter != nil && config.FieldsHexFilter.Match(name) {
+								v = hex.EncodeToString(b)
+							}
+						} else {
+							v = fmt.Sprintf("%v", v)
+						}
 					}
 				}
+
 				fields[name] = v
 			}
 		} else {
@@ -500,7 +511,7 @@ func (p *Parser) executeQuery(doc, selected dataNode, query string) (r interface
 				return nn.GetValue(), nil
 			}
 		}
-		// Fallback to get the string value representation
+
 		return iter.Current().Value(), nil
 	}
 

plugins/parsers/xpath/testcases/json_native_nonnested/expected.out

@@ -0,0 +1 @@
+foo a="a string",b=3.1415,c=true,d="map[d1:1 d2:foo d3:true d4:<nil>]",e="[master 42 true]",timestamp=1690193829 1690193829000000000

plugins/parsers/xpath/testcases/json_native_nonnested/telegraf.conf

@@ -0,0 +1,11 @@
+[[inputs.file]]
+  files = ["./testcases/json_string_representation/test.json"]
+  data_format = "xpath_json"
+
+  xpath_native_types = true
+
+  [[inputs.file.xpath]]
+    metric_name = "'foo'"
+    field_selection = "*"
+    timestamp = "timestamp"
+    timestamp_format = "unix"

plugins/parsers/xpath/testcases/json_native_nonnested/test.json

@@ -0,0 +1,13 @@
+{
+   "a": "a string",
+   "b": 3.1415,
+   "c": true,
+   "d": {
+     "d1": 1,
+     "d2": "foo",
+     "d3": true,
+     "d4": null
+   },
+   "e": ["master", 42, true],
+   "timestamp": 1690193829
+ }

plugins/parsers/xpath/testcases/json_string_representation/expected.out

@@ -0,0 +1 @@
+foo a="a string",b="3.1415",c="true",d="{\"d1\":1,\"d2\":\"foo\",\"d3\":true,\"d4\":null}",e="[\"master\",42,true]",timestamp="1690193829" 1690193829000000000

plugins/parsers/xpath/testcases/json_string_representation/telegraf.conf

@@ -0,0 +1,9 @@
+[[inputs.file]]
+  files = ["./testcases/json_string_representation/test.json"]
+  data_format = "xpath_json"
+
+  [[inputs.file.xpath]]
+    metric_name = "'foo'"
+    field_selection = "*"
+    timestamp = "timestamp"
+    timestamp_format = "unix"

plugins/parsers/xpath/testcases/json_string_representation/test.json

@@ -0,0 +1,13 @@
+{
+   "a": "a string",
+   "b": 3.1415,
+   "c": true,
+   "d": {
+     "d1": 1,
+     "d2": "foo",
+     "d3": true,
+     "d4": null
+   },
+   "e": ["master", 42, true],
+   "timestamp": 1690193829
+ }

plugins/parsers/xpath/testcases/name_expansion/telegraf.conf

@@ -6,5 +6,5 @@
   [[inputs.file.xpath]]
     metric_name = "'devices'"
     metric_selection = "/devices/*"
-    field_selection = "descendant::*"
+    field_selection = "descendant::*[not(*)]"
     field_name_expansion = true