Print the CSV header at -n 0, refuse --separator with a record format, refuse a date field in time(), read a bare number template as a format string, and say when no data is loaded

CHANGELOG.md

@@ -29,7 +29,8 @@ replacement, and each removed path, column or flag.
 - First release: the CLI, the library and the shipped data set.
 - The CLI renders exactly the template on its stdin, `echo '{/sv_SE.person} hihi' |
   fejkdata`, and adds nothing under `--format text`: `echo`'s newline ends the render, and `--separator`
-  defaults to empty. Whitespace around a JSON template is dropped. A lone reference,
+  defaults to empty. Whitespace around a JSON object, array or string template is
+  dropped, and a bare number or boolean is a format string, printed as written. A lone reference,
   `{/users}`, is the record `users` under `--format`. An argument other than a flag, or
   nothing piped in, exits 2; an empty or blank stdin is a template, and renders as it is.
 - Table categories: a category JSON naming a `rows` TSV beside it, with the options
@@ -100,7 +101,7 @@ replacement, and each removed path, column or flag.
   from a `sv_SE.birth-number` table under `sex`, in place of `sv_SE.ssn`, whose
   `ssn.mmdd`, `ssn.mmdd.m` and `ssn.mmdd.d` go with it. `en_US.ssn` now draws the
   ranges the SSA assigns and carries the columns `area`, `group` and `serial`, so
-  `--format csv en_US.ssn` writes a header where it used to fail; `en_US.itin` is new
+  `echo '{/en_US.ssn}' | fejkdata --format csv` writes a header where it used to fail; `en_US.itin` is new
   and carries the same three. `person.prefix` is null where a person has no title,
   where it used to be an empty string, so `--format sql` writes `NULL` and a `string`
   struct field reading it becomes `*string`.

README.md

@@ -30,8 +30,7 @@ born {date(1990-01-01,2010-12-31,'2006-01-02')}
 EOF
 ```
 
-fejkdata renders the template on its stdin: a format string, or a JSON object, array or
-string. Its `{…}` tokens reach the data by reference from the root,
+fejkdata renders the template on its stdin: a format string, or any JSON value. Its `{…}` tokens reach the data by reference from the root,
 `{/sv_SE.person.last}`, so shipped and `--data-path` categories are alike available,
 and everything outside them prints as written. A reference names a category, or a field
 inside one: each dot segment descends one level — folders, then the category (a JSON
@@ -40,9 +39,9 @@ these paths; a template reads one as `{/path}`.
 
 Under `--format text`, the default, what you send is what renders, and fejkdata adds nothing. A format string keeps every
 byte, so `echo`, a file or a heredoc ends each render with its newline, and `printf '%s'`
-sends none; `-n` joins renders with `--separator`, empty by default. A JSON template is
-any JSON value, and the whitespace around it is dropped; `42` or `true` alone renders its
-text, and `null` alone renders nothing. An empty or blank stdin is a template too, and
+sends none; `-n` joins renders with `--separator`, empty by default. A JSON object, array
+or string drops the whitespace around it, and `null` alone renders nothing; `42` or `true`
+alone is a format string, so it prints as written, newline included. An empty or blank stdin is a template too, and
 renders as it is. A JSON template that should end
 each render in a newline writes `\n` in its format, from a quoted heredoc or a file. A
 lone reference, one `{/…}` and nothing else, is the record it names under `--format`, one
@@ -55,8 +54,8 @@ template holding one from a quoted heredoc or a file. The rest of stdin's rules
 | `-d`, `--data-path D` | a directory to layer over the shipped data; repeatable, the last wins a name clash |
 | `--no-shipped-data` | load only the `--data-path` directories |
 | `-s`, `--seed N` | same seed, version and data: identical output |
-| `-n`, `--repeat N` | render the template N times (0 to 1048576), each an independent draw, streamed; `0` prints nothing, or `[]` under `--format json` |
-| `--separator S` | between repeated renders (default empty); a record `--format` ignores it |
+| `-n`, `--repeat N` | render the template N times (0 to 1048576), each an independent draw, streamed; `0` prints nothing, `[]` under `--format json`, or the header under `--format csv` |
+| `--separator S` | between repeated renders (default empty); refused with a record `--format`, which has its own separators |
 | `--format F` | `text` (default), `json`, `ndjson`, `csv` or `sql` — a record's columns, one record per row (json frames them as an array) |
 | `--table T` | the INSERT target for `--format sql`, ignored under any other (default: a lone reference's last segment, such as `users` for `{/users}`, else `records`) |
 | `--list` | print every path, then exit; `--repeat`, `--separator`, `--format` and `--table` do nothing here |
@@ -678,8 +677,10 @@ between the two days is reachable, so a layout with a clock draws the time too,
 `from` may equal `to`, which is that one day. The instant is UTC, so a zone in the
 layout prints `UTC` or `Z`.
 The quotes delimit a layout outside a selector only, so `[O'Fallon]` in an
-argument stays a name. A layout naming no field prints itself, `time` prints a date
-field as 1970-01-01, and `date` over one day with no clock field prints that day.
+argument stays a name. A layout naming no field prints itself, `date` with a clock-only
+layout prints the drawn instant's clock, and `date` over one day with no clock field prints
+that day; `time` with a layout naming a date field is refused, naming `date`, since `time`
+draws no date.
 Rejected at `New`: a bound that is no calendar date, or after the other, and an unquoted
 or double-quoted layout, naming the single-quoted one. `{seq()}` spans `Fake`
 calls and `repeat`, resets with a new generator, and is the natural primary key for

cmd/fejkdata/main.go

@@ -35,7 +35,7 @@ fejkdata <<'EOF'
 born {date(1990-01-01,2010-12-31,'2006-01-02')}
 EOF
 
-A template is a format string, or a JSON object, array or string. Its {…} tokens
+A template is a format string, or any JSON value. Its {…} tokens
 reach the data by reference from the root, {/sv_SE.person.last}, whether the data
 is shipped or layered with --data-path. A reference names a category or a dotted
 path into one ({/sv_SE.person.last}), a table's row by key or name
@@ -45,8 +45,9 @@ path into one ({/sv_SE.person.last}), a table's row by key or name
 Under --format text, the default, what you send is what renders, and fejkdata adds
 nothing: a format string keeps every byte, so echo's newline ends each render, and
 printf '%s' sends none. -n joins renders with --separator, empty by default. A JSON
-template is any JSON value, and the whitespace around it is dropped; 42 or true alone
-renders its text, and null nothing. An empty or blank stdin renders as it is. A
+object, array or string drops the whitespace around it, and null alone renders nothing;
+42 or true alone is a format string, printed as written. An empty or blank stdin
+renders as it is. A
 quoted heredoc, <<'EOF', passes $, a backslash or a quote as written.
 
 With --format json, ndjson, csv or sql the template must be a record — a JSON
@@ -61,9 +62,11 @@ one INSERT.
       --list             list the paths the data offers, each written after {/ in a template, then exit;
                          --repeat, --separator, --format and --table do nothing here
       --no-shipped-data  load only the --data-path directories
-  -n, --repeat N         render the template N times, 0..1048576 (default 1)
+  -n, --repeat N         render the template N times, 0..1048576 (default 1); 0 prints nothing,
+                         [] under --format json, or the header under --format csv
   -s, --seed N           same seed, version and data: identical output
-      --separator S      string between repeated renders (default empty); a record --format ignores it
+      --separator S      string between repeated renders (default empty); refused with a record
+                         --format, which has its own separators
       --table T          the INSERT target for --format sql, ignored under any other
                          (default: a lone reference's last segment, else records)
       --version          print the version, then exit
@@ -73,19 +76,20 @@ A short flag's value attaches or follows (-n3, -n 3); short flags bundle (-hn 3)
 `
 
 type invocation struct {
-	args      []string
-	dirs      []string
-	format    string
-	help      bool
-	list      bool
-	noShipped bool
-	repeat    int
-	seed      uint64
-	seeded    bool
-	separator string
-	table     string
-	tableSet  bool
-	version   bool
+	args         []string
+	dirs         []string
+	format       string
+	help         bool
+	list         bool
+	noShipped    bool
+	repeat       int
+	seed         uint64
+	seeded       bool
+	separator    string
+	separatorSet bool
+	table        string
+	tableSet     bool
+	version      bool
 }
 
 type flagDef struct {
@@ -117,7 +121,7 @@ var flagDefs = []flagDef{
 		in.seed, in.seeded = n, true
 		return nil
 	}},
-	{"separator", "", true, func(in *invocation, v string) error { in.separator = v; return nil }},
+	{"separator", "", true, func(in *invocation, v string) error { in.separator, in.separatorSet = v, true; return nil }},
 	{"table", "", true, func(in *invocation, v string) error { in.table, in.tableSet = v, true; return nil }},
 	{"version", "", false, func(in *invocation, _ string) error { in.version = true; return nil }},
 }
@@ -233,15 +237,16 @@ type recordFormat struct {
 	open   string
 	close  string
 	sep    string
+	writes string // what it writes between values, which --separator would seem to set
 }
 
 // recordFormats is every --format that writes records. json frames the records
 // as one array document; ndjson writes one object per line.
 var recordFormats = map[string]recordFormat{
-	"csv":    {header: (*fejkdata.Record).CSVHeader, line: func(r *fejkdata.Record, _ string) string { return r.CSVLine() }, sep: "\n"},
-	"json":   {line: jsonLine, open: "[", close: "]", sep: ",\n"},
-	"ndjson": {line: jsonLine, sep: "\n"},
-	"sql":    {line: func(r *fejkdata.Record, table string) string { return r.SQLInsert(table) }, sep: "\n"},
+	"csv":    {header: (*fejkdata.Record).CSVHeader, line: func(r *fejkdata.Record, _ string) string { return r.CSVLine() }, sep: "\n", writes: `always separates fields with ","`},
+	"json":   {line: jsonLine, open: "[", close: "]", sep: ",\n", writes: "writes the records as one JSON array"},
+	"ndjson": {line: jsonLine, sep: "\n", writes: "writes one record per line"},
+	"sql":    {line: func(r *fejkdata.Record, table string) string { return r.SQLInsert(table) }, sep: "\n", writes: "writes one INSERT per line"},
 }
 
 func jsonLine(r *fejkdata.Record, _ string) string { return r.JSON() }
@@ -268,6 +273,9 @@ func (in invocation) checkFlags() error {
 	if in.tableSet && in.table == "" && in.format == "sql" {
 		return errors.New("--table names the INSERT target, so it cannot be empty")
 	}
+	if f, records := recordFormats[in.format]; records && in.separatorSet && !in.list {
+		return fmt.Errorf("--separator joins text renders; --format %s %s", in.format, f.writes)
+	}
 	return nil
 }
 
@@ -400,16 +408,19 @@ func (in invocation) writeRecords(f *fejkdata.Generator, src input, w io.Writer)
 	return out.Flush()
 }
 
-// writeNoRecord writes zero records: an empty document where the format frames one, else nothing. It
-// draws one record and drops it, so a lone reference to nothing fails as it would with records.
+// writeNoRecord writes zero records: the header where the format writes one, an empty document
+// where it frames one, else nothing. It draws one record, which names the header's columns, so a
+// lone reference to nothing fails as it would with records.
 func writeNoRecord(record func() (*fejkdata.Record, error), format recordFormat, w io.Writer) error {
-	if _, err := record(); err != nil {
+	r, err := record()
+	switch {
+	case err != nil:
 		return err
+	case format.header != nil:
+		_, err = io.WriteString(w, format.header(r)+"\n")
+	case format.open != "":
+		_, err = io.WriteString(w, format.open+format.close+"\n")
 	}
-	if format.open == "" {
-		return nil
-	}
-	_, err := io.WriteString(w, format.open+format.close+"\n")
 	return err
 }
 
@@ -595,7 +606,7 @@ func run(args []string, stdin io.Reader, stdout, stderr io.Writer) int {
 		return 0
 	}
 	if err := in.write(f, src, stdout); err != nil {
-		return fail(stderr, err)
+		return fail(stderr, in.noDataCause(err))
 	}
 	return 0
 }
@@ -611,6 +622,18 @@ func readStdin(stdin io.Reader) (input, error) {
 	return parseInput(string(raw)), nil
 }
 
+// noDataCause adds the flags that left the generator without data to an error saying none is
+// loaded.
+func (in invocation) noDataCause(err error) error {
+	if !in.noShipped || len(in.dirs) > 0 || !strings.HasSuffix(err.Error(), "no data is loaded") {
+		return err
+	}
+	if te, isInput := err.(inputError); isInput {
+		return inputError{fmt.Errorf("%w: --no-shipped-data and no --data-path", te.error)}
+	}
+	return fmt.Errorf("%w: --no-shipped-data and no --data-path", err)
+}
+
 func fail(stderr io.Writer, err error) int {
 	var te inputError
 	if errors.As(err, &te) {

inline.go

@@ -1,6 +1,7 @@
 package fejkdata
 
 import (
+	"encoding/json"
 	"fmt"
 
 	"github.com/larvit/fejkdata/internal/grammar"
@@ -36,7 +37,7 @@ func (f *Generator) NewTemplate(input string) (*Template, error) {
 	f.mu.Lock()
 	defer f.mu.Unlock()
 	if err := loadInline(&f.root, n, "template", false); err != nil {
-		return nil, fmt.Errorf("fejkdata: %w", err)
+		return nil, f.noDataNote(fmt.Errorf("fejkdata: %w", err))
 	}
 	return &Template{g: f, n: n}, nil
 }
@@ -68,13 +69,18 @@ func compileInput(input string) (node, error) {
 	return compile(inputValue(input))
 }
 
-// inputValue reads an inline template as the value compile takes: the JSON value it holds, the
-// whitespace around it dropped, or the input itself as a format string when it is not JSON.
+// inputValue reads an inline template as the value compile takes: the JSON object, array, string
+// or null it holds, the whitespace around it dropped, or else the input itself as a format string,
+// a bare number or boolean among them.
 func inputValue(input string) jsonvalue.Value {
 	raw, err := jsonvalue.Decode([]byte(input))
 	if err != nil {
 		return jsonvalue.String(input)
 	}
+	switch raw.Any().(type) {
+	case json.Number, bool:
+		return jsonvalue.String(input)
+	}
 	return raw
 }
 

internal/builtinfunc/layout.go

@@ -83,10 +83,19 @@ func timeArg(a []string) error {
 	if err := layoutArity("time", 1, a); err != nil {
 		return err
 	}
-	_, err := layoutArg(a[0])
+	layout, err := layoutArg(a[0])
+	if err == nil && namesADateField(layout) {
+		return fmt.Errorf("time(layout): '%s' names a date field, and time draws no date; write date(from,to,layout)", layout)
+	}
 	return err
 }
 
+// namesADateField reports a layout that prints two instants differing in their date alone apart.
+func namesADateField(layout string) bool {
+	day, nextYear := time.Date(2001, 2, 3, 4, 5, 6, 0, time.UTC), time.Date(2010, 11, 12, 4, 5, 6, 0, time.UTC)
+	return day.Format(layout) != nextYear.Format(layout)
+}
+
 // datePrep draws a second in [from 00:00:00, to 23:59:59] UTC; the span is counted
 // in seconds, since a Duration overflows past 292 years.
 func datePrep(a []string) Call {

path.go

@@ -25,7 +25,7 @@ func childNamed(n node, seg string) (node, error) {
 		if child, ok := n.children[seg]; ok {
 			return child, nil
 		}
-		return nil, fmt.Errorf("no entry %q", seg)
+		return nil, noEntry{seg}
 	case *template:
 		if child, ok := n.fields[seg]; ok {
 			return child, nil

record.go

@@ -136,7 +136,7 @@ func (f *Generator) FakeRecord(path string) (*Record, error) {
 	f.loadShippedAt(segments)
 	_, n, tail, err := resolveCategory(f.root.children, segments)
 	if err != nil {
-		return nil, fmt.Errorf("fejkdata: %s: %w", path, err)
+		return nil, f.noDataNote(fmt.Errorf("fejkdata: %s: %w", path, err))
 	}
 	var frames frameStack
 	env := renderEnv{frames: &frames}

reference.go

@@ -1,6 +1,7 @@
 package fejkdata
 
 import (
+	"errors"
 	"fmt"
 	"strings"
 
@@ -77,7 +78,7 @@ func resolveCategory(root map[string]node, segments []string) (categorySegs []st
 	}
 	n, ok := g.children[segments[i]]
 	if !ok {
-		return nil, nil, nil, fmt.Errorf("no entry %q", segments[i])
+		return nil, nil, nil, noEntry{segments[i]}
 	}
 	return segments[:i+1], n, segments[i+1:], nil
 }
@@ -114,3 +115,18 @@ func refTokens(toks []grammar.Token) []string {
 	}
 	return refs
 }
+
+// noEntry is a path segment naming nothing in the folder it reads.
+type noEntry struct{ seg string }
+
+func (e noEntry) Error() string { return fmt.Sprintf("no entry %q", e.seg) }
+
+// noDataNote adds to err, where a segment names nothing and f holds no data at all, that no
+// data is loaded.
+func (f *Generator) noDataNote(err error) error {
+	var ne noEntry
+	if errors.As(err, &ne) && len(f.root.children) == 0 && len(f.root.unloaded) == 0 {
+		return fmt.Errorf("%w; no data is loaded", err)
+	}
+	return err
+}

render.go

@@ -27,7 +27,7 @@ func (f *Generator) Fake(path string) (string, error) {
 	f.loadShippedAt(segments)
 	n, pins, err := drawCallerPath(f.drawState, &f.root, segments)
 	if err != nil {
-		return "", fmt.Errorf("fejkdata: %s: %w", path, err)
+		return "", f.noDataNote(fmt.Errorf("fejkdata: %s: %w", path, err))
 	}
 	if _, ok := n.(*folder); ok {
 		return "", fmt.Errorf("fejkdata: %s names a folder, not a value", path)