Csv decoder (#1290)

* WIP: adding CSV decoder

* Adding CSV decoder

* Added CSV roundtrip

* Fixing from review

acceptance_tests/inputs-format.sh

@@ -3,6 +3,8 @@
 setUp() {
   rm test*.yml 2>/dev/null || true
   rm test*.properties 2>/dev/null || true
+  rm test*.csv 2>/dev/null || true
+  rm test*.tsv 2>/dev/null || true
   rm test*.xml 2>/dev/null || true
 }
 
@@ -40,6 +42,51 @@ EOM
   assertEquals "$expected" "$X"
 }
 
+testInputCSV() {
+  cat >test.csv <<EOL
+fruit,yumLevel
+apple,5
+banana,4
+EOL
+
+  read -r -d '' expected << EOM
+- fruit: apple
+  yumLevel: 5
+- fruit: banana
+  yumLevel: 4
+EOM
+
+  X=$(./yq e -p=csv test.csv)
+  assertEquals "$expected" "$X"
+
+  X=$(./yq ea -p=csv test.csv)
+  assertEquals "$expected" "$X"
+}
+
+testInputTSV() {
+  cat >test.tsv <<EOL
+fruit	yumLevel
+apple	5
+banana	4
+EOL
+
+  read -r -d '' expected << EOM
+- fruit: apple
+  yumLevel: 5
+- fruit: banana
+  yumLevel: 4
+EOM
+
+  X=$(./yq e -p=t test.tsv)
+  assertEquals "$expected" "$X"
+
+  X=$(./yq ea -p=t test.tsv)
+  assertEquals "$expected" "$X"
+}
+
+
+
+
 testInputXml() {
   cat >test.yml <<EOL
 <cat legs="4">BiBi</cat>

acceptance_tests/output-format.sh

@@ -102,6 +102,48 @@ EOM
   assertEquals "$expected" "$X"
 }
 
+testOutputCSV() {
+  cat >test.yml <<EOL
+- fruit: apple
+  yumLevel: 5
+- fruit: banana
+  yumLevel: 4
+EOL
+
+  read -r -d '' expected << EOM
+fruit,yumLevel
+apple,5
+banana,4
+EOM
+
+  X=$(./yq -o=c test.yml)
+  assertEquals "$expected" "$X"
+
+  X=$(./yq ea -o=csv test.yml)
+  assertEquals "$expected" "$X"
+}
+
+testOutputTSV() {
+  cat >test.yml <<EOL
+- fruit: apple
+  yumLevel: 5
+- fruit: banana
+  yumLevel: 4
+EOL
+
+  read -r -d '' expected << EOM
+fruit	yumLevel
+apple	5
+banana	4
+EOM
+
+  X=$(./yq -o=t test.yml)
+  assertEquals "$expected" "$X"
+
+  X=$(./yq ea -o=tsv test.yml)
+  assertEquals "$expected" "$X"
+}
+
 testOutputXml() {
   cat >test.yml <<EOL
 a: {b: {c: ["cat"]}}

cmd/utils.go

@@ -68,6 +68,10 @@ func configureDecoder() (yqlib.Decoder, error) {
 		return yqlib.NewPropertiesDecoder(), nil
 	case yqlib.JsonInputFormat:
 		return yqlib.NewJSONDecoder(), nil
+	case yqlib.CSVObjectInputFormat:
+		return yqlib.NewCSVObjectDecoder(','), nil
+	case yqlib.TSVObjectInputFormat:
+		return yqlib.NewCSVObjectDecoder('\t'), nil
 	}
 
 	return yqlib.NewYamlDecoder(), nil

examples/sample_objects.csv

@@ -0,0 +1,3 @@
+name,numberOfCats,likesApples,height
+Gary,1,true,168.8
+Samantha's Rabbit,2,false,-188.8

pkg/yqlib/csv_test.go

@@ -0,0 +1,273 @@
+package yqlib
+
+import (
+	"bufio"
+	"fmt"
+	"testing"
+
+	"github.com/mikefarah/yq/v4/test"
+)
+
+const csvSimple = `name,numberOfCats,likesApples,height
+Gary,1,true,168.8
+Samantha's Rabbit,2,false,-188.8
+`
+
+const expectedUpdatedSimpleCsv = `name,numberOfCats,likesApples,height
+Gary,3,true,168.8
+Samantha's Rabbit,2,false,-188.8
+`
+
+const csvSimpleShort = `Name,Number of Cats
+Gary,1
+Samantha's Rabbit,2
+`
+
+const tsvSimple = `name	numberOfCats	likesApples	height
+Gary	1	true	168.8
+Samantha's Rabbit	2	false	-188.8
+`
+
+const expectedYamlFromCSV = `- name: Gary
+  numberOfCats: 1
+  likesApples: true
+  height: 168.8
+- name: Samantha's Rabbit
+  numberOfCats: 2
+  likesApples: false
+  height: -188.8
+`
+
+const expectedYamlFromCSVMissingData = `- name: Gary
+  numberOfCats: 1
+  height: 168.8
+- name: Samantha's Rabbit
+  height: -188.8
+  likesApples: false
+`
+
+const csvSimpleMissingData = `name,numberOfCats,height
+Gary,1,168.8
+Samantha's Rabbit,,-188.8
+`
+
+const csvTestSimpleYaml = `- [i, like, csv]
+- [because, excel, is, cool]`
+
+const expectedSimpleCsv = `i,like,csv
+because,excel,is,cool
+`
+
+const tsvTestExpectedSimpleCsv = `i	like	csv
+because	excel	is	cool
+`
+
+var csvScenarios = []formatScenario{
+	{
+		description:  "Encode CSV simple",
+		input:        csvTestSimpleYaml,
+		expected:     expectedSimpleCsv,
+		scenarioType: "encode-csv",
+	},
+	{
+		description:  "Encode TSV simple",
+		input:        csvTestSimpleYaml,
+		expected:     tsvTestExpectedSimpleCsv,
+		scenarioType: "encode-tsv",
+	},
+	{
+		description:  "Encode Empty",
+		skipDoc:      true,
+		input:        `[]`,
+		expected:     "",
+		scenarioType: "encode-csv",
+	},
+	{
+		description:  "Comma in value",
+		skipDoc:      true,
+		input:        `["comma, in, value", things]`,
+		expected:     "\"comma, in, value\",things\n",
+		scenarioType: "encode-csv",
+	},
+	{
+		description:  "Encode array of objects to csv",
+		input:        expectedYamlFromCSV,
+		expected:     csvSimple,
+		scenarioType: "encode-csv",
+	},
+	{
+		description:    "Encode array of objects to custom csv format",
+		subdescription: "Add the header row manually, then the we convert each object into an array of values - resulting in an array of arrays. Pick the columns and call the header whatever you like.",
+		input:          expectedYamlFromCSV,
+		expected:       csvSimpleShort,
+		expression:     `[["Name", "Number of Cats"]] +  [.[] | [.name, .numberOfCats ]]`,
+		scenarioType:   "encode-csv",
+	},
+	{
+		description:    "Encode array of objects to csv - missing fields behaviour",
+		subdescription: "First entry is used to determine the headers, and it is missing 'likesApples', so it is not included in the csv. Second entry does not have 'numberOfCats' so that is blank",
+		input:          expectedYamlFromCSVMissingData,
+		expected:       csvSimpleMissingData,
+		scenarioType:   "encode-csv",
+	},
+	{
+		description:    "Parse CSV into an array of objects",
+		subdescription: "First row is assumed to be the header row.",
+		input:          csvSimple,
+		expected:       expectedYamlFromCSV,
+		scenarioType:   "decode-csv-object",
+	},
+	{
+		description:    "Parse TSV into an array of objects",
+		subdescription: "First row is assumed to be the header row.",
+		input:          tsvSimple,
+		expected:       expectedYamlFromCSV,
+		scenarioType:   "decode-tsv-object",
+	},
+	{
+		description:  "Round trip",
+		input:        csvSimple,
+		expected:     expectedUpdatedSimpleCsv,
+		expression:   `(.[] | select(.name == "Gary") | .numberOfCats) = 3`,
+		scenarioType: "roundtrip-csv",
+	},
+}
+
+func testCSVScenario(t *testing.T, s formatScenario) {
+	switch s.scenarioType {
+	case "encode-csv":
+		test.AssertResultWithContext(t, s.expected, processFormatScenario(s, NewYamlDecoder(), NewCsvEncoder(',')), s.description)
+	case "encode-tsv":
+		test.AssertResultWithContext(t, s.expected, processFormatScenario(s, NewYamlDecoder(), NewCsvEncoder('\t')), s.description)
+	case "decode-csv-object":
+		test.AssertResultWithContext(t, s.expected, processFormatScenario(s, NewCSVObjectDecoder(','), NewYamlEncoder(2, false, true, true)), s.description)
+	case "decode-tsv-object":
+		test.AssertResultWithContext(t, s.expected, processFormatScenario(s, NewCSVObjectDecoder('\t'), NewYamlEncoder(2, false, true, true)), s.description)
+	case "roundtrip-csv":
+		test.AssertResultWithContext(t, s.expected, processFormatScenario(s, NewCSVObjectDecoder(','), NewCsvEncoder(',')), s.description)
+	default:
+		panic(fmt.Sprintf("unhandled scenario type %q", s.scenarioType))
+	}
+}
+
+func documentCSVDecodeObjectScenario(t *testing.T, w *bufio.Writer, s formatScenario, formatType string) {
+	writeOrPanic(w, fmt.Sprintf("## %v\n", s.description))
+
+	if s.subdescription != "" {
+		writeOrPanic(w, s.subdescription)
+		writeOrPanic(w, "\n\n")
+	}
+
+	writeOrPanic(w, fmt.Sprintf("Given a sample.%v file of:\n", formatType))
+	writeOrPanic(w, fmt.Sprintf("```%v\n%v\n```\n", formatType, s.input))
+
+	writeOrPanic(w, "then\n")
+	writeOrPanic(w, fmt.Sprintf("```bash\nyq -p=%v sample.%v\n```\n", formatType, formatType))
+	writeOrPanic(w, "will output\n")
+
+	separator := ','
+	if formatType == "tsv" {
+		separator = '\t'
+	}
+
+	writeOrPanic(w, fmt.Sprintf("```yaml\n%v```\n\n",
+		processFormatScenario(s, NewCSVObjectDecoder(separator), NewYamlEncoder(s.indent, false, true, true))),
+	)
+}
+
+func documentCSVEncodeScenario(w *bufio.Writer, s formatScenario, formatType string) {
+	writeOrPanic(w, fmt.Sprintf("## %v\n", s.description))
+
+	if s.subdescription != "" {
+		writeOrPanic(w, s.subdescription)
+		writeOrPanic(w, "\n\n")
+	}
+
+	writeOrPanic(w, "Given a sample.yml file of:\n")
+	writeOrPanic(w, fmt.Sprintf("```yaml\n%v\n```\n", s.input))
+
+	writeOrPanic(w, "then\n")
+
+	expression := s.expression
+
+	if expression != "" {
+		writeOrPanic(w, fmt.Sprintf("```bash\nyq -o=%v '%v' sample.yml\n```\n", formatType, expression))
+	} else {
+		writeOrPanic(w, fmt.Sprintf("```bash\nyq -o=%v sample.yml\n```\n", formatType))
+	}
+	writeOrPanic(w, "will output\n")
+
+	separator := ','
+	if formatType == "tsv" {
+		separator = '\t'
+	}
+
+	writeOrPanic(w, fmt.Sprintf("```%v\n%v```\n\n", formatType,
+		processFormatScenario(s, NewYamlDecoder(), NewCsvEncoder(separator))),
+	)
+}
+
+func documentCSVRoundTripScenario(w *bufio.Writer, s formatScenario, formatType string) {
+	writeOrPanic(w, fmt.Sprintf("## %v\n", s.description))
+
+	if s.subdescription != "" {
+		writeOrPanic(w, s.subdescription)
+		writeOrPanic(w, "\n\n")
+	}
+
+	writeOrPanic(w, fmt.Sprintf("Given a sample.%v file of:\n", formatType))
+	writeOrPanic(w, fmt.Sprintf("```%v\n%v\n```\n", formatType, s.input))
+
+	writeOrPanic(w, "then\n")
+
+	expression := s.expression
+
+	if expression != "" {
+		writeOrPanic(w, fmt.Sprintf("```bash\nyq -p=%v -o=%v '%v' sample.%v\n```\n", formatType, formatType, expression, formatType))
+	} else {
+		writeOrPanic(w, fmt.Sprintf("```bash\nyq -p=%v -o=%v sample.%v\n```\n", formatType, formatType, formatType))
+	}
+	writeOrPanic(w, "will output\n")
+
+	separator := ','
+	if formatType == "tsv" {
+		separator = '\t'
+	}
+
+	writeOrPanic(w, fmt.Sprintf("```%v\n%v```\n\n", formatType,
+		processFormatScenario(s, NewCSVObjectDecoder(separator), NewCsvEncoder(separator))),
+	)
+}
+
+func documentCSVScenario(t *testing.T, w *bufio.Writer, i interface{}) {
+	s := i.(formatScenario)
+	if s.skipDoc {
+		return
+	}
+	switch s.scenarioType {
+	case "encode-csv":
+		documentCSVEncodeScenario(w, s, "csv")
+	case "encode-tsv":
+		documentCSVEncodeScenario(w, s, "tsv")
+	case "decode-csv-object":
+		documentCSVDecodeObjectScenario(t, w, s, "csv")
+	case "decode-tsv-object":
+		documentCSVDecodeObjectScenario(t, w, s, "tsv")
+	case "roundtrip-csv":
+		documentCSVRoundTripScenario(w, s, "csv")
+
+	default:
+		panic(fmt.Sprintf("unhandled scenario type %q", s.scenarioType))
+	}
+}
+
+func TestCSVScenarios(t *testing.T) {
+	for _, tt := range csvScenarios {
+		testCSVScenario(t, tt)
+	}
+	genericScenarios := make([]interface{}, len(csvScenarios))
+	for i, s := range csvScenarios {
+		genericScenarios[i] = s
+	}
+	documentScenarios(t, "usage", "csv-tsv", genericScenarios, documentCSVScenario)
+}