§9.5.1

Blank and Spacer Rows

An empty row—meaningful structure or just spacing?

⚖️ Equivalent formsTwo renderings conform. A document uses one of them throughout.📄 Page ruleCan be applied from the page alone.Read §9.5.1 in the RFC ↗

The requirement

A genuinely printed blank row inside a table MAY be rendered as an empty row or left out; both forms conform, and a document MUST use one of them throughout. A blank row that is the only thing separating two groups of rows SHOULD be rendered as an empty row.

When it applies

Sketch of the printed page — layout, not output

| ROU assets total   | 10,234 |
|                    |        |   ← printed blank row: on this page it is the
| Lease liabilities: |        |     only section boundary — rendered as an empty row

Reading the examples. Each one pairs a source page with the output the rule expects. ✅ Sample: the rule applies. ❌ Counter-sample: it looks like it applies and does not, or a rendering got it wrong. A non-conforming output, in red, is shown only for contrast. … marks content left out of an excerpt.

When the blank row is load-bearing

On a page with no rules, no bands and no repeated headers, the whitespace is the structure. Dropping a blank row there merges two sections the page deliberately kept apart — as in a lease note where a single blank line is all that separates the right-of-use totals from "Lease liabilities:".

A blank row whose removal would change how a reader groups the rows around it SHOULD be rendered as a row of empty cells, in its printed position.
A purely decorative blank line — even spacing between every row, padding at the foot of a block — carries no such meaning and MAY be left out.
An empty row MUST NOT be rendered where the page prints none: the row is kept because it is printed, never added to make groups easier to read.

✅ SampleThe rule applies

A genuinely printed blank row separates the 00xx class codes from the 01xx codes`ert`

A genuinely printed blank row separates the 00xx class codes from the 01xx codes, in every block; it is kept as a row of empty cells. The four side-by-side CLASS CODE | ELR | D RATIO blocks continue one class-code series, so each block is its own table (see Column Continuation (§9.2.2)); the first is shown.

Source page

Expected output · table 1 of 4

CLASS CODE	ELR	D RATIO
0010	0.912	0.42
0015	0.955	0.41
0025	0.972	0.38
0040	1.164	0.40

0100	0.888	0.39
0115	1.010	0.41
0130	0.842	0.37
0145	1.075	0.42

✅ SampleThe rule applies

In the long-cruise timetable`lakecruise`

In the long-cruise timetable, a fully empty row is printed between the outbound run ending at Seeblick ○ and the return run starting at Seeblick; it is the only boundary between the two direction segments and is kept. Stop rows printed in heavier type are bold cell by cell. The header symbol under each service number is ⊛.

Source page

Expected output · table 1 of 2, excerpt around the blank row

	Täglich Daily
Kurs / Service	101/102	103/104	105/106	107/108	109/110	113/114	117/118
	⊛	⊛	⊛	⊛	⊛	⊛	⊛
Lindenhafen	09.20	10.20	11.20	12.20	13.20	15.20	17.20	Lindenhafen
Aurenbach	09.47	↓	↓	12.47	13.47	↓	↓	Aurenbach
…
Ufeninsel	11.19	12.19	13.19	14.19	15.19	17.19	19.29	Ufeninsel
Seeblick ○	11.30	12.30	13.30	14.30	15.30	17.30	19.40	○ Seeblick

Seeblick	11.40	12.40	13.40	14.40	15.40	17.40	19.45	Seeblick
Ufeninsel	11.53	12.51	13.51	14.53	15.53	17.51	19.56	Ufeninsel
…

✅ SampleThe rule applies

A thin printed empty row sits between the 12:05–35 block and the 13:00 block`2012_tushin_nittei`

A thin printed empty row sits between the 12:05–35 block and the 13:00 block, the only thing separating the morning heats from the afternoon semi-finals; it is kept.

Source page

Expected output · excerpt around the blank row

時刻	トラック	跳躍（高さ）	跳躍（距離）	投擲
９：３０	女共４×１００ｍ予５組１着＋３
５０	男共４×１００ｍ予６組０着＋８
１０：２０	男３０００ｍ決２組タイムレース			女砲丸投
５５	男８００ｍ予３組２着＋２	女走高跳Ｂ
１１：１０	女１５００ｍ決２組タイムレース
３５	女１００ｍＨ予４組３着＋４
５０	男２年１００ｍ予５組３着＋１
１２：０５	女２年１００ｍ予６組２着＋４
２０	男１年１００ｍ予６組２着＋４		男三段跳
３５	女１年１００ｍ予５組３着＋１

１３：００	女１００ｍＨ準２組４着
…

❌ Counter-sampleLooks like it applies and does not — or a rendering got it wrong

Blank rows are the only separation between the Experience Period`e22`

Blank rows are the only separation between the Experience Period, Annual … Trend and Projected groups; they are kept. A rendering that drops them runs the three groups together; it still conforms, but draws a warning. The red ($8.75) already carries its sign in the parentheses, so the colour is decoration and is not rendered.

Source page

Expected output · excerpt around the first two blank rows

Rx Claim Information	Generic	Brand	Specialty	Total
Experience Period Scripts / 1000	9,882	1,094	89	11,065
…
Experience Period Paid PMPM	$9.36	$17.90	$29.51	$56.77
Experience Period Rx Rebates PMPM				($8.75)

Annual Util Trend	1.031	1.003	1.070	1.028
Annual Unit Cost Trend	0.986	1.138	1.077	1.088
Annual Allowed Trend	1.016	1.141	1.152	1.119
Annual Deductible Trend	1.013	1.013	1.013	1.035
Annual Paid Trend	1.003	1.164	1.154	1.134
Annual Paid Trend Net of Rebates	n/a	n/a	n/a	1.127
Months of Trend	24	24	24	24

Projected Scripts / 1000	10,502	1,099	102	11,704
…

❌ Counter-sampleLooks like it applies and does not — or a rendering got it wrong

Printed blank lines separate the cohort blocks and set off the Average Issue Age line`3samehead`

The bold cohort bands already separate the blocks, so these spacers MAY be dropped; they are kept here. The rendering below drops them, and it also records 10.9% for the 2004+ share of premium where the page prints 32031.9%, and splits the Issue Year band into a label cell and empty cells although the page prints one full-width band. Each ruled Total row is a total, so its label cell is a <th> and the stub carries levels (all 0: nothing is indented).

Source page

Expected output · excerpt

	Count	% of Count	Premium	% of Premium
Issue Year
1999	80	3.5%	217,072	3.2%
2000	316	13.7%	964,763	14.2%
2001	374	16.2%	1,091,387	16.0%
2002	733	31.8%	2,110,680	31.0%
2003	570	24.7%	1,695,230	24.9%
2004+	234	10.1%	738,975	32031.9%
Total	2,307	100.0%	6,818,108	100.0%

Issue Age
<30	4	0.2%	4,451	0.1%
…
80+	3	0.1%	8,175	0.1%
Total	2,307	100.0%	6,818,108	100.0%

Average Issue Age	60.6

Attained Age
…

Non-conforming output

What the rule rules out.

	Count	% of Count	Premium	% of Premium
Issue Year
1999	80	3.5%	217,072	3.2%
2000	316	13.7%	964,763	14.2%
2001	374	16.2%	1,091,387	16.0%
2002	733	31.8%	2,110,680	31.0%
2003	570	24.7%	1,695,230	24.9%
2004+	234	10.1%	738,975	10.9%
…

❌ Counter-sampleLooks like it applies and does not — or a rendering got it wrong

Printed gaps sit under the BUILDINGS / OCCUPIED BY OWNER band, around the BUS`3cell`

PERS. PROP. band and between Group 5 and Group 6; the last is the only thing splitting the rate groups in two, so it is kept, and the others are kept with it. The BUS. PERS. PROP. band prints its label in its own ruled cell with one empty cell across the rest of the row, so it is not a full-width span. The trailing gap under Group 10 is padding and is not shown.

Source page

Expected output · table 1 of 2, excerpt

CLASSIFICATION	FRAME			JOISTED MASONRY			LT. NON-COMBUSTIBLE			FIRE RESISTIVE			500K / GL LIMIT
	1-6	7-8	9-10	1-6	7-8	9-10	1-6	7-8	9-10	1-6	7-8	9-10	500K / GL LIMIT
BUILDINGS OCCUPIED BY OWNER

Store Pac	13.33	16.57	19.37	9.59	11.86	13.83	7.51	9.29	10.84	4.89	5.82	6.68	N/A

BUS. PERS. PROP.

Rate Group 1	35.37	37.63	39.27	33.30	35.47	36.98	32.69	34.76	36.28	29.21	31.10	32.42	.94
Group 2	37.91	40.33	42.09	35.73	38.00	39.65	35.03	37.28	38.90	31.31	33.29	34.74	.97
Group 3	39.73	42.25	44.08	37.40	39.81	41.52	36.71	39.06	40.73	32.79	34.87	36.40	1.51
Group 4	41.22	43.84	45.74	38.81	41.28	43.10	38.06	40.50	42.25	34.03	36.18	37.75	1.62
Group 5	41.86	44.52	46.44	39.41	41.92	43.76	38.65	41.12	42.93	34.51	36.72	38.32	1.51

Group 6	43.23	45.98	47.99	40.72	43.31	45.19	39.93	42.49	44.32	35.67	37.96	39.61	1.52
…

Encoding an empty region

The encoding of an empty region is free. Its existence is not.

An empty region inside a table body MAY be encoded as individual empty <td> cells or as one blank cell spanning it (colspan / rowspan): three leading empty <td>s per row and one colspan="3" rowspan="8" blank block render identically, and both conform. Where the page draws no separation inside the empty region, neither encoding is preferred.
Carve-out: in a loss triangle or similar matrix, every empty position MUST be its own empty cell, never part of a span — each is a distinct, meaningful absence. See Triangles and Matrix Tables (§9.8.2).
The half that is not free: the cells and columns MUST exist. A column proven only by its header still exists (Barely-Existing Columns and Headers (§9.3.1)), and every value in it MUST stay reachable. Encoding is a choice; deletion is an error.
Blank space beside a header label is not free: it MUST be rendered as separate empty cells, not as an extended span, unless the page prints a genuinely merged cell. See Barely-Existing Columns and Headers (§9.3.1).

Rule id tables-blank-and-spacer-rows · tables either rows disputed

When it applies

When the blank row is load-bearing

Encoding an empty region

Neighbouring rules, and how they differ

More