Method / Metric definitions · 口徑對照
Metric definitions
同一個名嘅香港數字，官方往往有兩套口徑。呢頁講清楚邊個數同邊個數唔可以相加、唔可以互換， 以及引用時要指明邊一份刊物。
繁體中文
錯用統計最常見嘅方式，唔係引錯數，而係把兩套口徑當成同一條序列：例如將「陸上非住院人口」 當成「比年中人口少咗十萬人」，或者將酒店同賓館房間相加做一個官方從來冇公布過嘅「住宿供應」。
呢頁列出 33 組口徑對照（33 組兩邊都已經係已登記 fact），涉及 75 條已登記 fact。每一組都會寫明：兩邊嘅計算範圍、唔包括咩、值同時點由邊條 fact 派生、以及出處網址。若某組仲有未核實嘅第三個口徑（例如只寫咗喺 fact 註記），會喺該組下面嘅 〈未齊全嘅部分〉照實講清楚，唔會靠估補一個數。
問答入口同樣受呢一層保護：問題命中某組口徑嘅觸發詞（每組下面列出）時，Auto Chat 會喺答案下面即刻提示「呢條數有兩套官方口徑」並連返呢頁。 要把未轉正嘅口徑「轉正」，次序係先在 統計館藏 登記來源同 fact， 再喺呢頁加側邊——口徑層本身永遠唔會自創數字。
值一律唔會手寫：呢頁每個數字都係由 統計館藏已登記嘅官方 fact 即時派生，所以口徑層永遠唔會出現官方 fact 以外嘅數。 機器可讀版本：/api/v1/definitions.json； 全部收錄規則見 編輯方法。
English
The most common way to misuse Hong Kong statistics is not quoting the wrong number, but treating two official bases as one series — for example reading the land-based non-institutional population as “a hundred thousand people fewer than the mid-year estimate”, or adding hotel and guesthouse rooms into an “accommodation supply” figure that no official source publishes.
This page lists 33 calibre entries (33 with both sides already registered as facts), drawing on 75 registered facts. Each entry states the scope of both sides, what is excluded, and the fact, reference date and source behind every figure. Where a third, unverified basis exists (for example one recorded only in a fact note) it is spelled out under “Gap” rather than filled in with an estimate.
The chat entry point is covered too: when a question matches an entry’s trigger words (listed under each entry), Auto Chat immediately flags “this measure has more than one official basis” and links back here. Promoting a pending basis follows the same order — register the source and the fact in thestatistics collection first, then add the side here; this layer never invents a number.
No figure here is typed by hand. Every value is derived at render time from the sourced facts in the statistics collection, so this layer cannot contain a number that is not in a registered fact. Machine-readable version: /api/v1/definitions.json; editorial rules: method.
兩套官方口徑 / Two official bases
「香港有幾大」應該答陸地面積，抑或連海域一齊計？
When someone asks how large Hong Kong is, does the answer cover land only, or land together with sea?
陸地總面積同海洋總面積係兩套分開公布嘅官方數字。相加就會得出一個官方從來冇公布過嘅「總面積」，所以引用時一定要講明係陸地定連海域。
Total land area and total sea area are published as separate official figures. Adding them creates a territorial total that no official source has ever published, so a citation must say whether it is land only or land plus sea.
1 114.6 平方公里 / 1,114.6 km²
只計陸地 / Land only
陸地總面積 · asOf 2025-12 / Total land area
注意：唔包括海域；官方並冇公布「陸地加海域」嘅單一總數 / Note: Sea is excluded; no single combined land-plus-sea total is published
備註：年底數字。香港政府一站通「香港概況」寫作 1,114.57 平方公里。
出處：香港統計數字一覽 2026年版 · 相關專題
約 1 640 平方公里 / About 1,640 km²
只計海域 / Sea only
海洋總面積 · asOf 2025-07-31 / Total sea area
注意：同陸地總面積分開計算，唔可以相加當成「總面積」 / Note: Counted separately from land area; the two must not be added into a single territorial total
備註：漁護署海岸公園概述。與陸地總面積是兩套數字，不可相加當成「總面積」而不讀原文。
出處：認識海岸公園及海岸保護區 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：香港面積、陸地面積、海洋面積、香港有幾大 / Triggered in chat by: how large is hong kong, land area, sea area
兩套官方口徑 / Two official bases
「香港有幾多人」，點解會有兩個官方數字？
Why does “how many people live in Hong Kong” have two official answers?
兩個數字係兩條唔同嘅官方序列：一條係全港年中人口估計，另一條係按區議會分區統計嘅陸上非住院人口。方法同覆蓋範圍唔同，相差唔可以解讀成「人口少咗」或者「流失」。
The two figures belong to different official series: the territory-wide mid-year population estimate, and the land-based non-institutional population used for district statistics. Coverage differs, so the gap must not be read as population loss.
7 498 900 / 7,498,900
全港年中人口估計（官方人口序列） / Territory-wide mid-year population estimate (the official population series)
2025年年中人口 · asOf 2025-06 / Mid-year population 2025
7 400 500 / 7,400,500
陸上非住院人口（區議會分區統計口徑） / Land-based non-institutional population (the district-statistics base)
陸上非住院人口 · asOf 2025 / Land-based non-institutional population
注意：唔可以同全港年中人口相比，當成「比年中人口少」 / Note: Must not be compared with the mid-year estimate as though the difference were a headcount shortfall
備註：綜合住戶統計調查口徑，與全港年中人口估計（7 498 900）不同，不可直接當成「少了約 10 萬人」。
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：香港人口、香港有幾多人、幾多人、陸上非住院人口、年中人口 / Triggered in chat by: hong kong population, population of hong kong, how many people live in hong kong
兩套官方口徑 / Two official bases
「幾多香港人講英文」——慣用同能說係兩件事？
How many people in Hong Kong use English? Speaking ability and usual language differ.
同一份人口普查，按「慣用交談語言」統計同按「能說」統計係兩套完全唔同嘅數字。前者講日常主要語言，後者講能力，唔可以互相代替。
In the same census, statistics by usual spoken language and by ability to speak are two different measures: one describes the everyday main language, the other a capability. They cannot stand in for each other.
4.6% / 4.6%
慣用英語（日常交談語言） / Usual language English (everyday spoken language)
慣用英語（5歲及以上） · asOf 2021 / Usual language: English (aged 5+)
注意：唔等於「能說英語」 / Note: Not the same as the ability to speak English
57.7%（2021，較2011年45.1%上升） / 57.7% (2021, up from 45.1% in 2011)
能說英語（語言能力） / Able to speak English (ability)
能說英語的人口比例 · asOf 2021 / Proportion able to speak English
注意：唔等於日常慣用英語 / Note: Not the same as using English as a usual language
備註：人口普查語言使用情況：能說英文比率，與「慣用英語」4.6%是兩套口徑。
出處：2021年人口普查：語言使用情況 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：幾多人講英文、慣用英語、能說英語 / Triggered in chat by: usual language english, able to speak english
兩套官方口徑 / Two official bases
「幾多香港人講普通話」——慣用同能說差幾遠？
How many people in Hong Kong use Putonghua? Usual language versus ability.
同一份人口普查，按「慣用交談語言」統計嘅普通話人口同按「能說」統計嘅數字相差好遠：前者係日常主要語言，後者係能力。
In the same census the Putonghua figures by usual spoken language and by ability differ widely: one is a daily main language, the other a capability.
2.3% / 2.3%
慣用普通話（日常交談語言） / Usual language Putonghua (everyday spoken language)
慣用普通話（5歲及以上） · asOf 2021 / Usual language: Putonghua (aged 5+)
注意：唔等於「能說普通話」 / Note: Not the same as the ability to speak Putonghua
56.5%（2021，較2011年49.5%上升） / 56.5% (2021, up from 49.5% in 2011)
能說普通話（語言能力） / Able to speak Putonghua (ability)
能說普通話的人口比例 · asOf 2021 / Proportion able to speak Putonghua
注意：唔等於日常慣用普通話 / Note: Not the same as using Putonghua as a usual language
備註：人口普查：能說普通話比率，比慣用普通話2.3%高得多，反映「能力」與「慣用」之分。
出處：2021年人口普查：語言使用情況 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：幾多人講普通話、慣用普通話、能說普通話 / Triggered in chat by: usual language putonghua, able to speak putonghua
兩套官方口徑 / Two official bases
「香港一年有幾多日假期」，點解兩個數字都有人講？
Why do two different counts circulate for holidays in Hong Kong?
憲報公布嘅公眾假期同《僱傭條例》嘅法定假日係兩份唔同名單、兩套數目。講「假期」要講清楚係邊一份，尤其係計僱員待遇嘅時候。
General holidays as gazetted and statutory holidays under the Employment Ordinance are different lists with different counts. Any statement about holidays must say which list is meant, especially for employment entitlements.
17 日，另加每個星期日 / 17 named days, plus every Sunday
憲報公布嘅公眾假期 / General holidays as gazetted
2026年公眾假期（憲報公布的節日） · asOf 2026 / Gazetted general holidays in 2026
注意：同法定假日唔係同一份名單 / Note: Not the same list as statutory holidays
備註：公眾假期與法定假日不是同一份名單。
15 日 / 15 days
《僱傭條例》法定假日 / Statutory holidays under the Employment Ordinance
2026年法定假日 · asOf 2026 / Statutory holidays in 2026
注意：數目會隨年份逐步增加，唔可以當成固定常數 / Note: The count rises over coming years, so it is not a fixed constant
備註：2026年起新增復活節星期一。將逐步增加至2030年的17日。
出處：2026年的法定假日 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：公眾假期、法定假日、幾多日假期 / Triggered in chat by: public holiday, statutory holiday, how many holidays
兩套官方口徑 / Two official bases
「香港有幾多張病床」——點解兩個官方數字相差幾千張？
Why do official hospital-bed figures for Hong Kong differ by thousands?
兩套口徑量度嘅設施範圍唔同：一套包含醫院管理局、私家醫院及懲教院所嘅病床；另一套用經合組織定義，剔除急症室觀察病床、日間病床及育嬰病床。做國際比較要用後者，講本地設施總量要用前者。
The two bases cover different facilities: one counts beds in Hospital Authority hospitals, private hospitals and correctional institutions; the OECD definition excludes accident-and-emergency observation, day and nursery beds. Use the second for international comparison and the first for local capacity.
37 142（每千人 4.9，比率為臨時數字） / 37,142 (4.9 per 1,000; rate provisional)
醫院管理局、私家醫院及懲教院所合共 / Hospital Authority, private hospitals and correctional institutions combined
病床（醫管局、私家醫院及懲教院所） · asOf 2025-12 / Hospital beds (HA, private and correctional)
備註：包括所有在醫院管理局轄下醫院、私家醫院及懲教院所轄下醫院內所設的病床。由2019年起，私家醫院病床包括住院及日間病床。
出處：香港統計數字一覽 2026年版 · 相關專題
32 725（每千人 4.4，比率為臨時數字） / 32,725 (4.4 per 1,000; rate provisional)
經合組織定義（國際比較用） / OECD definition (for international comparison)
病床（經合組織定義） · asOf 2025-12 / Hospital beds (OECD definition)
注意：唔包括急症室觀察病床、日間病床及育嬰病床 / Note: Excludes accident-and-emergency observation, day and nursery beds
備註：只包括醫管局及私家醫院病床，不包括急症室觀察病床、日間病床及育嬰病床。
出處：香港統計數字一覽 2026年版 · 相關專題
30,824 張 / 30,824
醫院管理局單獨公布嘅整體病床數（另一個統計時點） / Hospital Authority only, its own published overall bed count (a different reference date)
公立醫院病床數（2025年3月底） · asOf 2025 / Public hospital beds (end-March 2025)
注意：唔包括私家醫院及懲教院所 / Note: Private hospitals and correctional institutions are not included
備註：醫院管理局《主要績效指標No.66》：截至2025年3月31日整體病床數約30,824張。
出處：醫院管理局：主要績效指標（病床數） · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：病床、床位、幾多張病床 / Triggered in chat by: hospital beds, beds per
兩套官方口徑 / Two official bases
「港鐵有幾長」，點解兩個官方數字相差幾十公里？
Why do two official figures for MTR network length differ by tens of kilometres?
兩個數字都係官方公布：一個係整個重鐵系統嘅全包口徑，另一個只計本地重鐵路線。講「港鐵網絡」時要講明包邊啲系統。
Both figures are official: one is the all-in heavy-rail system figure, the other covers local heavy rail only. State which systems are included when quoting a network length.
271 公里 / 271 km
整個重鐵系統（全包口徑） / The whole heavy-rail system (all-in)
港鐵路網營運里程 · asOf 2025年6月 / MTR network operating length
備註：港鐵公司中期／年報公佈。
約 237 公里（本地重鐵） / about 237 km (local heavy rail)
本地重鐵路線 / Local heavy rail only
港鐵本地重鐵路線總長度 · asOf 2025 / MTR local heavy rail network length
注意：不包括輕鐵及高鐵香港段 / Note: Light rail and the high-speed rail Hong Kong section are excluded
備註：港鐵本地重鐵網絡約 237 公里（不包括約 36 公里輕鐵及約 26 公里高鐵香港段）；運輸及物流局另有「整個重鐵系統約 271 公里」之全包口徑。
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：港鐵有幾長、港鐵里程、路網里程、重鐵 / Triggered in chat by: mtr network length, mtr length, heavy rail
兩套官方口徑 / Two official bases
「限制區上落車證明書」嘅申請聯絡機構，究竟有幾多間？
How many contact organisations handle applications for the restricted-zone certificate?
同一份官方指南內有兩個數目：第四章正文叫讀者透過社聯統籌嘅聯絡機構申請，同一段又叫人參閱表二；表二逐列編號嘅總數同正文講嘅數目唔一致。兩個都係官方原文，文件冇交代差額原因，所以引用時要講明用邊個，唔可以當同一個數、更加唔可以相加減。
One official guide contains two counts: the fourth chapter sends applicants to the contact organisations coordinated by the HKCSS and in the same breath points to the guide’s second table, whose numbered rows add up to a different total. Both are official and the difference is left unexplained, so a citation must say which one it uses — treating them as one figure, or combining them, would create a number no official source has published.
80 間聯絡機構 / 80 contact organisations
第四章正文講嘅數目（社聯統籌） / As stated in the fourth chapter (coordinated by the HKCSS)
「限制區上落車證明書」聯絡機構數目（官方正文） · asOf 2026-06 / Restricted-zone certificate contact organisations (guide text)
注意：冇逐間列出機構；同表二列數唔一致 / Note: Does not name the organisations, and does not match the number of numbered rows in the guide’s second table
備註：運輸署《殘疾人士公共交通指南》2025 年版（2026 年 6 月修訂本）第四章正文：屬指定類別嘅殘疾人士可向香港社會服務聯會統籌的 80 間聯絡機構申請「殘疾乘客在限制區上落車證明書」。同一段叫讀者參閱表二，而表二逐列編號只到 76（見另一條 fact）——兩個數目都係官方原文，文件冇解釋差額，所以兩條並列、唔相加減（口徑分歧見 /calibre）。
76 間（表二編號一至七十六） / 76 organisations (Table 2 numbered 1 to 76)
表二逐列編號嘅總數（中文頁同英文頁逐列核對） / Total of the numbered rows in the guide’s second table (checked row by row against the facing English page)
「限制區上落車證明書」聯絡機構數目（官方表二逐列） · asOf 2026-06 / Restricted-zone certificate contact organisations (Table 2 rows)
注意：唔包括官方清單未列嘅機構；亦唔可以當成正文所講嘅數目 / Note: Excludes organisations the official table does not list, and is not the figure given in the text
備註：同一份官方指南第四章表二：中文頁逐列編號由 1 到 76，每列有名稱、電話同服務地區；對面英文頁係同一批機構嘅英文名（兩頁逐列核對過，唔一致嘅列已逐列註明）。呢個數同正文講嘅 80 間並列，唔互換、唔相加。
未齊全嘅部分：差額原因官方文件冇交代：表二冇講有機構退出、抑或正文數目包含其他渠道，所以本站唔會自行補一個解釋，亦唔會自行把兩個數調和。 / Gap: The official document does not explain the difference — the guide’s second table does not say whether organisations have left the scheme, nor whether the text figure covers other channels — so this site offers no invented explanation and does not reconcile the two numbers.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：限制區上落車證明書、聯絡機構、傷健協會 / Triggered in chat by: restricted zone certificate, contact organisations, hkcss
兩套官方口徑 / Two official bases
郊野公園面積，要唔要加埋特別地區？
Does the country park area figure include special areas?
兩個官方數字只差在「特別地區」。要留意部分特別地區本來就喺郊野公園範圍之內，已經計入公園面積，所以加總時只可以加公園範圍外嘅部分。
The two official figures differ by special areas. Some special areas lie inside country parks and are already inside the park total, so only the part outside the parks may be added.
44 842 公頃 / 44,842 hectares
郊野公園連特別地區（官方公布嘅合併數） / Country parks together with special areas (the published combined figure)
郊野公園及特別地區面積 · asOf 2024-03-01 / Area of country parks and special areas
備註：其中郊野公園合計 43 997 公頃。香港政府一站通寫「約佔土地 40%」，面積以漁護署表為準。
出處：認識郊野公園及海岸公園 · 相關專題
43 997 公頃 / 43,997 hectares
只計郊野公園本身 / Country parks only, before the special-areas add-back
郊野公園面積（不計算特別地區加成） · asOf 2024-03-01 / Country parks area only (before special-areas add-back)
注意：唔包括公園範圍外嘅特別地區 / Note: Special areas outside the parks are not included
備註：25 個郊野公園合計。22 個特別地區中，有 1 152 公頃在公園範圍之內，只有公園範圍外的 845 公頃另加。
出處：認識郊野公園及海岸公園 · 相關專題
43,997 公頃（郊野公園），連特別地區合共 44,842 公頃 / 43,997 ha (country parks); 44,842 ha incl. special areas
漁護署另一份公布同時列出兩邊數字（唔係第三個口徑） / Another AFCD publication listing both figures side by side (not a third basis)
郊野公園面積（2024年） · asOf 2024 / Country parks area (2024)
注意：唔可以當成「郊野公園以外另一個面積」；佢一條 fact 就包住上面兩個口徑 / Note: Must not be read as a further area figure: this single entry carries both of the bases above
備註：漁農自然護理署：郊野公園面積 43,997 公頃，如連同特別地區共 44,842 公頃（現有 country-parks-area 為另一口徑，此為總計）。
出處：漁農自然護理署：郊野公園數目及面積 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：郊野公園面積、郊野公園有幾大、特別地區 / Triggered in chat by: country park area, special area
同一指標、多個官方出口 / Same indicator, several official outlets
同一時點嘅註冊醫生數目，點解兩個政府統計處刊物寫得唔同？
Why do two Census and Statistics Department publications give different counts of registered doctors for the same date?
同一時點、兩個官方刊物列出唔同數目。要引用就一定要指名係邊一份刊物，唔好當成「香港有兩個互相矛盾嘅事實」。
For the same reference date two official publications list different counts. A citation must name the publication; this is not two contradictory facts about Hong Kong.
16 905 / 16,905
政府統計處《香港統計數字一覽》（只計正式註冊） / Census and Statistics Department, Hong Kong in Figures (full registration only)
註冊醫生 · asOf 2025-12 / Registered doctors
注意：唔包括以有限度註冊受僱嘅醫生 / Note: Doctors registered under limited registration for employment are excluded
備註：《香港統計數字一覽》註明：只計列於普通及非本地居民名冊嘅正式註冊醫生；唔包括根據《醫務註冊條例》以有限度註冊受僱嘅醫生（例如大學研究或教學、豁免診所執業等聘任）。
出處：香港統計數字一覽 2026年版 · 相關專題
17,809 名（每千人 2.4） / 17,809 (2.4 per 1,000 population)
政府統計處《香港統計年刊》及衞生署（截至年底註冊醫生數目） / Census and Statistics Department, Hong Kong Annual Digest of Statistics, with the Department of Health (registered doctors as at year end)
註冊醫生人數（2025年） · asOf 2025 / Registered doctors (2025)
備註：統計處《香港統計年刊》及衞生署：截至年底註冊醫生數目。
出處：政府統計處：註冊醫護專業人員數目 · 相關專題
未齊全嘅部分：兩邊嘅名冊範圍仲未核齊：《香港統計數字一覽》嘅註腳已經核實（只計正式註冊、唔包括有限度註冊），但統計處網頁表嘅註腳只交代年底數字同執業證書續期安排，冇寫名冊範圍；差額嘅成因（例如另一邊有冇計有限度註冊）未喺刊物頁面證實，本站唔會推測，亦唔會擅自選一個當正確答案。 / Gap: Coverage is only half verified: the footnote of Hong Kong in Figures has been checked (full registration only, limited registration excluded), while the C&SD web table’s note only covers the year-end reference date and practising-certificate renewal and says nothing about scope. The cause of the difference (for example whether the other side counts limited registration) has not been established on a publication page, so the site neither guesses nor picks one number as the sole truth.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：註冊醫生、醫生人數、幾多醫生 / Triggered in chat by: registered doctors, number of doctors
兩套官方口徑 / Two official bases
「香港有幾多護士」——點解兩個官方數字相差一萬幾千人？
Why do two official nurse counts for Hong Kong differ by more than ten thousand?
兩套口徑嘅專業範圍唔同：一套只計註冊護士，另一套把註冊護士同登記護士一齊計。兩個數字唔可以互換，亦唔可以相加；引用時要講明係「只計註冊護士」定「註冊加登記」。
The two bases cover different professional groups: one counts registered nurses only, the other counts registered and enrolled nurses together. The figures are not interchangeable and must not be added; a citation has to say whether it is registered nurses alone or registered plus enrolled.
74 767 / 74,767
政府統計處《香港統計數字一覽》（只計註冊護士） / Census and Statistics Department, Hong Kong in Figures (registered nurses only)
註冊護士 · asOf 2025-12 / Registered nurses
注意：唔包括登記護士 / Note: Enrolled nurses are not included
出處：香港統計數字一覽 2026年版 · 相關專題
91,416 名（每千人 12.3） / 91,416 (12.3 per 1,000 population)
政府統計處《香港統計年刊》及衞生署（註冊護士＋登記護士） / Census and Statistics Department, Hong Kong Annual Digest of Statistics, with the Department of Health (registered and enrolled nurses)
註冊護士及登記護士人數（2025年） · asOf 2025 / Registered and enrolled nurses (2025)
備註：統計處《香港統計年刊》及衞生署：截至年底註冊護士及登記護士數目。
出處：政府統計處：註冊醫護專業人員數目 · 相關專題
未齊全嘅部分：登記護士嘅單獨數目未立 fact，所以兩邊嘅差額只可以由標籤理解，本站唔會把差額寫成「登記護士人數」呢個官方數字。另一邊《香港統計數字一覽》嘅護士註腳未逐字核對（醫生嗰項已核），名冊範圍以標籤為準。 / Gap: The number of enrolled nurses on its own is not registered as a fact, so the difference between the two sides is explained by their labels only and is never presented here as an official “enrolled nurses” count. The nurse footnote in Hong Kong in Figures has not been checked word for word (the doctors entry has), so the register scope rests on the labels.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：護士人數、註冊護士、登記護士、幾多護士 / Triggered in chat by: nurses, registered nurses, enrolled nurses
同一指標、多個官方出口 / Same indicator, several official outlets
同一年的訪港旅客數字，點解三個官方出口寫法都唔同？
Why do three official outlets print the same year’s visitor arrivals differently?
同一年、同一個指標，由旅遊發展局、政府統計處同香港便覽分別公布；數值一致，差別只在進位同單位寫法。三個唔係三個唔同統計，亦唔互相矛盾。
The same year and indicator are published by the Tourism Board, the Census and Statistics Department and the Hong Kong Fact Sheets. The values agree; only rounding and units differ. These are not three rival statistics.
49,894,832 人次 / 49,894,832 arrivals
按實際人次公布嘅版本 / Published as an unrounded headcount
訪港旅客總人次（2025年） · asOf 2025 / Visitor arrivals (2025)
備註：旅發局：2025年全年訪港旅客49.9百萬人次（按年+12%）；其中內地約37.8百萬（佔76%）。
出處：香港旅遊發展局：2025全年訪港旅客統計 · 相關專題
4 989.5 萬人次 / 49.895 million
以「萬人次」公布嘅版本 / Published in units of ten thousand arrivals
訪港旅客 · asOf 2025 / Visitor arrivals
出處：香港統計數字一覽 2026年版 · 相關專題
4 990 萬人次（內地 3 780 萬、非內地 1 210 萬） / 49.9 million (Mainland 37.8M, non-Mainland 12.1M)
香港便覽〈旅遊業〉嘅版本（另按客源地分列） / Hong Kong Fact Sheet on Tourism, with a breakdown by source market
訪港旅客（2025年） · asOf 2025 / Visitor arrivals (2025)
備註：香港便覽〈旅遊業〉：按客源地劃分嘅訪港旅客人次。
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：訪港旅客、旅客人次、幾多旅客 / Triggered in chat by: visitor arrivals, tourist arrivals
唔可以相加 / Not additive
「幾多人嚟香港」同「幾多旅客嚟香港」係唔係同一件事？
Are inbound passenger movements and visitor arrivals the same thing?
出入境客流係所有經口岸出入嘅人次，包含回港居民；訪港旅客只計以香港為目的地嘅旅客。兩個數字唔可以互換或者相減。
Inbound passenger traffic counts every arrival through a control point, including returning residents; visitor arrivals count only visitors whose destination is Hong Kong. The two figures cannot be swapped or subtracted.
1.673 億人次 / 167.304 million
出入境客流（陸路、空路、海路合計） / Inbound passenger traffic (land, air and sea combined)
抵港旅客（按交通模式合計） · asOf 2025 / Passenger arrivals by all modes
注意：包含回港居民，唔係訪港旅客 / Note: Includes returning residents; this is not visitor arrivals
備註：陸路 1.403 億、空路 2 295 萬、海路 406 萬。此為出入境客流，包含回港居民，不是訪港旅客。
出處：香港統計數字一覽 2026年版 · 相關專題
4 989.5 萬人次 / 49.895 million
訪港旅客（以香港為目的地） / Visitor arrivals (Hong Kong as the destination)
訪港旅客 · asOf 2025 / Visitor arrivals
出處：香港統計數字一覽 2026年版 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：入境客流、出入境人次、幾多人嚟香港 / Triggered in chat by: inbound passenger, passenger traffic
唔可以相加 / Not additive
酒店同賓館房間可唔可以加埋做一個「住宿供應」數字？
Can hotel and guesthouse rooms be added into one accommodation-supply figure?
酒店同賓館係兩套牌照制度，房間面積、價格同客群都唔同。官方分開公布，加埋一齊就會得出官方冇公布過嘅數。
Hotels and guesthouses operate under two separate licensing regimes, with different room sizes, prices and clientele. They are published separately, and adding them creates a figure no official source publishes.
333 間、93 472 個房間 / 333 hotels, 93,472 rooms
持牌酒店 / Licensed hotels
持牌酒店 · asOf 2025-12 / Licensed hotels
注意：唔可以同賓館房間相加做「住宿供應」 / Note: Must not be added to guesthouse rooms as “accommodation supply”
備註：民政事務總署牌照事務處數字。全年入住率 87%。賓館是另一套牌照。
出處：香港統計數字一覽 2026年版 · 相關專題
1 234 間、11 463 個房間 / 1,234 guesthouses, 11,463 rooms
持牌賓館 / Licensed guesthouses
持牌賓館 · asOf 2025-12 / Licensed guesthouses
注意：同酒店係兩套牌照，唔可以相加 / Note: A separate licence class from hotels; the two must not be added
備註：民政事務總署牌照事務處；賓館與酒店是兩套牌照，全年入住率 68%。不可與酒店房間相加作「住宿供應」。
出處：香港統計數字一覽 2026年版 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：酒店房間、賓館房間、住宿供應 / Triggered in chat by: hotel rooms, guesthouse rooms, accommodation supply
唔可以相加 / Not additive
「香港有幾多菲律賓人／外籍人士」——入境處同人口普查係唔係同一件事？
How many Filipinos or foreign nationals are in Hong Kong — do the Immigration Department and the census measure the same thing?
入境處數字係按出入境紀錄點算某一日身在香港嘅某國籍人士；人口普查係按種族自我認同統計。兩者時點、身份定義同計法都唔同，唔可以相加、相減，亦唔可以當成互補——官方從來冇公布過「外籍人士總數」呢個合計。
The Immigration Department figure counts people of a given nationality present on one particular day, from arrival and departure records; the census counts ethnic identity as self-reported. Reference date, definition of status and method all differ, so the two cannot be added, subtracted or treated as complementary — no official total of foreign nationals is published.
195 574 人 / 195,574
入境處按出入境點算（某一日身在香港） / Immigration Department count (present in Hong Kong on one day)
身在香港的菲律賓國籍人士 · asOf 2025-12 / Philippine nationals present in Hong Kong
注意：唔等於持有居留身分嘅外籍居民總數，亦唔可以同普查種族分類相加 / Note: Not the resident foreign-national population, and must not be added to the census ethnic classification
備註：入境處按出入境紀錄點算特定一天身在香港的該國籍人士，不是持有居留身分的外籍居民總數，亦不可與人口普查種族分類相加。
182 215 人 / 182,215
入境處按出入境點算（印度尼西亞籍，某一日身在香港） / Immigration Department count (Indonesian nationals present on one day)
身在香港的印度尼西亞國籍人士 · asOf 2025-12 / Indonesian nationals present in Hong Kong
注意：唔等於持有居留身分嘅外籍居民總數，亦唔可以同普查種族分類相加 / Note: Not the resident foreign-national population, and must not be added to the census ethnic classification
備註：入境處按出入境紀錄點算特定一天身在香港的該國籍人士，不是持有居留身分的外籍居民總數，亦不可與人口普查種族分類相加。
39 587 人 / 39,587
入境處按出入境點算（印度籍，某一日身在香港） / Immigration Department count (Indian nationals present on one day)
身在香港的印度國籍人士 · asOf 2025-12 / Indian nationals present in Hong Kong
注意：唔等於持有居留身分嘅外籍居民總數，亦唔可以同普查種族分類相加 / Note: Not the resident foreign-national population, and must not be added to the census ethnic classification
備註：入境處按出入境紀錄點算特定一天身在香港的該國籍人士，不是持有居留身分的外籍居民總數，亦不可與人口普查種族分類相加。
619 568 人（約佔人口 8.4%） / 619,568 (about 8.4% of population)
人口普查種族自我認同（非華裔少數族裔） / Census ethnic self-identification (non-Chinese ethnic minorities)
少數族裔人口（2021年人口普查） · asOf 2021 / Ethnic minority population (2021 Census)
注意：同入境處按國籍點算嘅人數唔同時點、唔同定義，唔可以相加或相減 / Note: A different reference date and definition from the Immigration count; the two must not be added or subtracted
備註：統計處2021人口普查《主題性報告：少數族裔人士》：按種族自我認同，非華裔少數族裔共619,568人。
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：菲律賓籍、菲律賓人、印尼籍、印度籍、外籍人士、種族分類、少數族裔 / Triggered in chat by: filipino, indonesian, indian, ethnic minorities, nationality count
唔可以相加 / Not additive
失業人數加就業不足人數，等唔等於「冇全職工作嘅人」？
Do unemployed plus underemployed people equal everyone without full-time work?
失業同就業不足係兩套定義、兩組數字。相加唔等於「冇全職工作嘅總人數」，官方亦從來冇公布過嗰個總數。
Unemployment and underemployment are two definitions producing two sets of figures. Adding them does not give the number of people without full-time work, a total no official source publishes.
139 000 / 139,000
失業人數 / Unemployed persons
失業人數 · asOf 2025 / Unemployed persons
注意：唔可以同就業不足相加 / Note: Must not be added to the underemployed
備註：刊物以千人計為 139。失業率與失業人數是同一套綜合住戶統計調查。
出處：香港統計數字一覽 2026年版 · 相關專題
1.5%（56 000 人） / 1.5% (56,000 persons)
就業不足人數 / Underemployed persons
就業不足率 · asOf 2025 / Underemployment rate
注意：唔可以同失業相加成「冇全職工作嘅總人數」 / Note: Must not be added to the unemployed to form a “no full-time work” total
備註：就業不足與失業是兩套數字，不可相加當成「沒有全職工作的總人數」。
出處：香港統計數字一覽 2026年版 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：失業、就業不足 / Triggered in chat by: unemployment, underemployment
唔可以相加 / Not additive
「終審法院有幾多案件」——上訴許可申請加實質上訴，等唔等於案件總數？
How many cases does the Court of Final Appeal have — do leave applications plus substantive appeals give a total?
兩個數字係同一批案件嘅前後階段（先申請上訴許可，獲批之後才進入實質上訴），唔係兩類互不相干嘅案件。相加會重複計算，官方亦冇公布過嗰個總數。
The two figures are sequential stages of the same cases — leave must be granted before an appeal is heard — not two disjoint groups. Adding them double-counts, and no official total is published.
229 宗（2024） / 229 (2024)
上訴許可申請（案件入口） / Leave applications (case intake)
終審法院上訴許可申請 · asOf 2024 / Court of Final Appeal leave applications
注意：唔可以同實質上訴相加成「終審法院案件總數」 / Note: Must not be added to substantive appeals to form a “total CFA caseload”
備註：司法機構公佈附表 A 案件入口分項；同一批案件嘅實質上訴階段另見 fact 條目 jud-cfa-substantive-appeals-2024，兩者唔可以相加當總數。
出處：香港政府新聞公報：各級法院案件量統計 · 相關專題
24 宗（2024） / 24 (2024)
實質上訴（許可獲批之後嘅階段） / Substantive appeals (the stage after leave is granted)
終審法院實質上訴案件 · asOf 2024 / Court of Final Appeal substantive appeals
注意：同一批案件嘅後續階段，唔可以同許可申請相加 / Note: A later stage of the same cases; must not be added to leave applications
備註：司法機構公佈附表 A 分項，屬上訴許可獲批之後嘅階段；唔可以同上訴許可申請相加當「終審法院案件總數」。
出處：香港政府新聞公報：各級法院案件量統計 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：終審法院、上訴許可申請、實質上訴 / Triggered in chat by: court of final appeal, leave to appeal, substantive appeal
兩套官方口徑 / Two official bases
同一批上訴法庭刑事上訴，點解司法機構年報同政府新聞公報寫唔同數？
Why do the Judiciary annual report and the government press release give different counts for criminal appeals in the Court of Appeal?
兩個都係官方數字，但計法唔同：年報「刑事上訴」只計來自原訟法庭同區域法院嘅上訴；公報附表 A 同一行另外併入判刑覆核同新增嘅「其他案件」（律政司轉介／保留法律問題）。本站按年報逐項分項相加核對，差額同嗰兩行完全對得上。兩個數字唔可以互換，亦唔可以相加。
Both figures are official, but they are counted differently: the annual report’s criminal appeals line covers appeals from the Court of First Instance and the District Court only, while the same line in Annex A of the press release also folds in sentence reviews and the new “other cases” category (Secretary for Justice’s reference or reservation of a question of law). Adding up the annual report’s line items reproduces the difference exactly. The two figures are not interchangeable and must not be added.
刑事上訴 272 宗、民事上訴 577 宗（2024） / 272 criminal + 577 civil appeals (2024)
司法機構年報上訴法庭分項（同一條 fact 另載民事上訴，兩者分開列） / Court of Appeal line item in the Judiciary annual report (the same fact also carries the civil appeals, listed separately)
高等法院上訴法庭收案 · asOf 2024 / High Court, Court of Appeal caseload
注意：唔包括判刑覆核同「其他案件」 / Note: Sentence reviews and “other cases” are not included
備註：司法機構年報上訴法庭分項，只計來自原訟法庭及區域法院嘅刑事上訴；同一條 fact 另載民事上訴。同一份年報另有判刑覆核同「其他案件」分項。
277 宗（2024） / 277 (2024)
政府新聞公報附表 A 同一行（已併入判刑覆核同「其他案件」） / The same line in Annex A of the government press release (sentence reviews and “other cases” folded in)
高等法院上訴法庭刑事上訴（新聞公報附表 A） · asOf 2024 / High Court Court of Appeal criminal appeals (press release Annex A)
注意：範圍比年報分項闊，唔可以當年報數字用 / Note: Broader than the annual report line; must not be used as the annual report figure
備註：政府新聞公報附表 A 同一行，已併入判刑覆核同「其他案件」（律政司轉介／保留法律問題）；範圍比年報上訴法庭刑事上訴分項闊，唔可以當年報數字用。
出處：香港政府新聞公報：各級法院案件量統計 · 相關專題
未齊全嘅部分：公報原文冇逐項列明「刑事上訴」一行包含邊幾類案件；上面嘅對應關係係本站按年報分項相加核對得出（差額逐項對得上），未喺公報原文搵到逐項說明。 / Gap: The press release itself does not itemise what its criminal appeals line contains; the correspondence stated above is reconstructed by adding up the annual report’s line items, which reproduces the difference exactly, and is not spelt out item by item in the release.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：上訴法庭刑事上訴、上訴法庭收案、高等法院上訴法庭案件 / Triggered in chat by: court of appeal criminal appeals, court of appeal caseload
兩套官方口徑 / Two official bases
同一批原訟法庭民事案件，點解司法機構年報同政府新聞公報寫唔同數？
Why do the Judiciary annual report and the government press release give different counts for civil matters in the Court of First Instance?
年報「民事管轄權」計原訟法庭自己處理嘅民事案件（包括破產同公司清盤訴訟），唔計來自審裁處嘅民事上訴；公報附表 A 同一行把嗰批上訴併入去，所以兩個官方數唔同。本站按年報逐項分項相加核對，差額同來自審裁處嘅民事上訴完全對得上。兩者唔可以互換，亦唔可以相加。
The annual report’s civil jurisdiction line counts civil business handled by the Court of First Instance itself (including bankruptcy and companies winding-up matters) but not civil appeals from tribunals; the same line in Annex A of the press release folds those appeals in, so the two official figures differ. Adding up the annual report’s line items reproduces the difference exactly. They are not interchangeable and must not be added.
刑事 2,162 宗、民事 20,001 宗（2024） / 2,162 criminal + 20,001 civil matters (2024)
司法機構年報原訟法庭民事管轄權分項（同一條 fact 另載刑事分項，兩者分開列） / Civil jurisdiction line item in the Judiciary annual report (the same fact also carries the criminal items, listed separately)
高等法院原訟法庭收案 · asOf 2024 / High Court, Court of First Instance caseload
注意：唔包括來自審裁處嘅民事上訴 / Note: Civil appeals from tribunals are not included
備註：司法機構年報原訟法庭分項：民事部分係民事管轄權，包括破產及公司清盤訴訟，唔包括來自審裁處嘅民事上訴；刑事部分係刑事案、機密雜項程序同刑事雜項程序相加。
20 126 宗（2024） / 20 126 (2024)
政府新聞公報附表 A 同一行（已併入來自審裁處嘅民事上訴） / The same line in Annex A of the government press release (civil appeals from tribunals folded in)
高等法院原訟法庭民事案件（新聞公報附表 A） · asOf 2024 / High Court Court of First Instance civil matters (press release Annex A)
注意：範圍比年報分項闊，唔可以當年報數字用，亦唔可以同刑事分項相加 / Note: Broader than the annual report line; must not be used as the annual report figure or added to the criminal items
備註：政府新聞公報附表 A 同一行，已併入來自審裁處嘅民事上訴；範圍比年報原訟法庭民事管轄權分項闊，唔可以當年報數字用。
出處：香港政府新聞公報：各級法院案件量統計 · 相關專題
未齊全嘅部分：公報原文冇逐項列明「civil jurisdiction」一行包含邊幾類案件；上面嘅對應關係係本站按年報分項相加核對得出（差額逐項對得上），未喺公報原文搵到逐項說明。 / Gap: The press release itself does not itemise what its civil jurisdiction line contains; the correspondence stated above is reconstructed by adding up the annual report’s line items, which reproduces the difference exactly, and is not spelt out item by item in the release.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：原訟法庭民事、原訟法庭收案、高等法院原訟法庭案件 / Triggered in chat by: court of first instance civil, court of first instance caseload
兩套官方口徑 / Two official bases
同一年度嘅醫療衞生總開支佔本地生產總值，點解官方有兩個百分比？
Why are there two official percentages for total health expenditure as a share of GDP in the same financial year?
兩個都出自《本地醫療衞生總開支帳目》，但一個係連新冠相關開支一齊計嘅總開支口徑，另一個係撇除新冠相關開支之後嘅口徑。兩個數字唔可以互換，亦唔可以相加；引用時要講明有冇撇除新冠相關開支。
Both come from the domestic health accounts, but one is the total expenditure basis including COVID related expenditure while the other removes it. The two figures are not interchangeable and must not be added; a citation has to say whether COVID related expenditure is included.
10.0%（2022-23 含新冠） / 10.0% (FY2022-23 incl. COVID)
總開支口徑（包括新冠相關開支） / Total expenditure basis (including COVID related expenditure)
醫療衞生總開支佔GDP · asOf 2022-23 / Total health expenditure as share of GDP
注意：唔可以當成撇除新冠之後嘅數字 / Note: Must not be used as the figure net of COVID related expenditure
備註：《本地醫療衞生總開支帳目》總開支口徑（包括2019冠狀病毒病相關開支）；同一指標撇除新冠相關開支之後嘅官方數字，另見 fact 條目 health-gdp-2022-23-excl-covid。
出處：醫務衞生局：本地醫療衞生總開支帳目 · 相關專題
8.4%（2022-23 撇除新冠） / 8.4% (FY2022-23, excl. COVID)
撇除新冠相關開支之後嘅口徑 / Basis after removing COVID related expenditure
醫療衞生總開支佔GDP（撇除新冠相關開支） · asOf 2022-23 / Total health expenditure as share of GDP (excl. COVID-19 related expenditure)
注意：唔包括新冠相關開支，亦唔可以同總開支口徑相加 / Note: Excludes COVID related expenditure and must not be added to the total-expenditure basis
備註：《香港年報2024》引述《本地醫療衞生總開支帳目》：撇除2019冠狀病毒病相關開支後，二零二二至二三年度醫療衞生總開支佔本地生產總值嘅比重。同一頁另有公共醫療衞生總開支佔醫療衞生總開支嘅比重，屬另一個指標。
出處：《香港年報2024》第十九章：醫務衞生 · 相關專題
未齊全嘅部分：同一份刊物另有「公共醫療衞生總開支佔醫療衞生總開支嘅比重」，係另一個指標（分母唔係本地生產總值），本站未為佢立法 fact；舊註記一度寫過一個百分比，但喺年報同帳目頁都核實唔到，所以已經移除，唔會列出亦唔靠估。 / Gap: The same publication also carries the public share of total health expenditure, which is a different indicator (its denominator is not GDP) and is not registered as a fact here; an earlier note carried a percentage that could not be verified on the yearbook or accounts pages, so it has been removed rather than asserted.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：醫療衞生總開支、醫療開支佔、醫療衞生開支 / Triggered in chat by: health expenditure, health spending share
同一指標、多個官方出口 / Same indicator, several official outlets
性別比率應該引邊條官方序列？
Which official series should a sex ratio be quoted from?
同一指標、三條唔同官方序列：綜合住戶統計調查口徑、按區議會分區統計嘅陸上非住院人口口徑，同人口估計口徑。覆蓋範圍唔同，時點亦唔完全對齊（人口估計口徑嘅官方比率行，最新刊出嘅一年係前一年），所以引用時唔只要寫時點，仲要寫係邊一條序列；三者唔可以互相代替。
One indicator, three official series: the General Household Survey base, the land-based non-institutional population base used for district statistics, and the population-estimate base. Coverage differs and the time points do not fully line up (the latest published year of the population-estimate ratio line is the preceding year), so a citation needs both the date and the series; the three are not interchangeable.
每千名女性對 828 名男性 / 828 males per 1,000 females
綜合住戶統計調查口徑（全港） / General Household Survey base (territory-wide)
性別比率（2025年） · asOf 2025 / Sex ratio (2025)
備註：政府統計處綜合住戶統計調查 2025 年全港性別比率為 828（每千名女性對應男性數目）；人口估計口徑嘅官方比率行另見 fact 條目 pop-mid-sex-ratio-2024（年刊最新刊出嘅一年）。
每千名女性對 828 名男性 / 828 males per 1,000 females
陸上非住院人口口徑（按區議會分區統計） / Land-based non-institutional population base (district statistics)
性別比率 · asOf 2025 / Sex ratio
注意：唔包括住院人士、水上人口同流動居民，唔可以同人口估計口徑互換 / Note: Excludes institutional, marine and mobile residents; not interchangeable with the population-estimate base
每千名女性對 832 名男性（2024） / 832 males per 1,000 females (2024)
人口估計口徑（年中人口；官方比率行最新刊出嘅一年） / Population-estimate base (mid-year population; latest year published for the official ratio line)
性別比率（人口估計口徑） · asOf 2024 / Sex ratio (population-estimate basis)
注意：時點同另外兩條唔同，唔可以當同一年嘅數字比較，亦唔可以互換 / Note: Its reference year differs from the other two; not to be read as the same year’s figure or swapped with them
備註：《香港統計年刊》2025 年版表 1.1「性別比率」：男性數目相對每千名女性嘅比率，口徑為年中人口（人口估計）；該表最新一年為二零二四年，所以同綜合住戶統計調查口徑嘅年份唔相同。
未齊全嘅部分：三條序列嘅時點唔齊：綜合住戶統計調查口徑同陸上非住院人口口徑有同一年嘅數字，但人口估計口徑嘅官方比率行最新只刊到前一年（新一期年刊未出），所以三者唔可以在同一時點直接比較。要同一年對比就等新一期年刊，唔准用除法自行推算一個比率出嚟。 / Gap: The three series do not line up in time: the General Household Survey base and the land-based non-institutional population base are published for the same year, but the official ratio line on the population-estimate base has only been published up to the preceding year (the next digest is not out yet), so the three cannot be compared at one time point. A same-year comparison must wait for the next digest; deriving a ratio by division is not allowed.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：性別比率、男女比例 / Triggered in chat by: sex ratio, males per
同一指標、多個官方出口 / Same indicator, several official outlets
同一年嘅本地生產總值，政府網站同統計處刊物係唔係同一個數？
Do the government website and the statistical publication give the same GDP for the same year?
政府網站摘要同統計處刊物會列出略有差異嘅數字，刊物亦會標明日後或會修訂。兩個數字都係官方，唔可以互換，亦唔可以相加；引用時要指名睇邊一份，並保留「或會修訂」嘅標示。
The government website summary and the statistical publication list slightly different figures, and the publication marks its figure as revisable. Both are official and must not be swapped or added; name the source you quote and keep that flag.
33,298 億港元 / HK$3,329.8 billion
香港政府一站通頁面數字 / Figure on the GovHK page
本地生產總值（政府網站數字） · asOf 2025 / Gross domestic product (government website figure)
備註：香港政府一站通 2026 年 7 月頁面數字。同年統計處刊物列出嘅本地生產總值見 fact 條目 gdp-2025-censtatd（另一官方出口）；兩個數字唔可以互換，亦唔可以相加，引用時要指名係邊一份。
33,318 億港元 / HK$3,331.8 billion
政府統計處刊物數字（標示或會修訂） / Statistical publication figure (marked as revisable)
本地生產總值（統計處刊物數字） · asOf 2025 / Gross domestic product (statistical publication)
注意：唔可以同政府網站數字相加或互換 / Note: Must not be added to, or swapped with, the government website figure
備註：《香港統計數字一覽2026年版》以當時市價計算；刊物標示日後或會修訂。政府網站數字另見 fact 條目 gdp-2025-govhk，兩者唔可以互換或相加。
出處：香港統計數字一覽 2026年版 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：本地生產總值、gdp / Triggered in chat by: gross domestic product, gdp
兩套官方口徑 / Two official bases
「財赤幾多」——要唔要計發債？
Should a government deficit figure include bond issuance?
同一份財政文件有兩個口徑：計入發行及償還政府債券之後，同未計發債之前。兩個數字唔可以互換，亦唔可以相加，講「赤字」時要講明係邊個口徑。
The same fiscal document presents two bases: after government bond issuance and redemption, and before it. The two figures are not interchangeable and must not be added, so state the basis when quoting a deficit.
802.68 億港元（2024-25） / HK$80.27bn (FY2024-25)
計入發行及償還政府債券之後 / After government bond issuance and redemption
政府綜合赤字（計入發債後） · asOf 2024-25 / Consolidated government deficit (after bond issuance)
備註：計入發行及償還政府債券後嘅口徑。同一份財政文件嘅計發債前口徑另見 fact 條目 treasury-deficit-prebond-2024-25；兩個口徑唔可以互換，亦唔可以相加，講「赤字」要講明係邊個口徑。
出處：財經事務及庫務局：政府財務數據 · 相關專題
1,883.11 億港元（2024-25，計發債前） / HK$188.311bn (FY2024-25, before bond issuance)
計發債之前 / Before government bond issuance
政府綜合赤字（計發債前） · asOf 2024-25 / Consolidated government deficit (before bond issuance)
注意：唔可以同計發債後嘅口徑互換或相加 / Note: Must not be swapped with, or added to, the after-bond basis
備註：同一份財政文件嘅計發債前口徑；計入發行及償還政府債券後嘅數字另見 fact 條目 treasury-deficit-2024-25。兩個口徑唔可以互換或相加。
出處：財經事務及庫務局：政府財務數據 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：財赤、赤字、財政差額 / Triggered in chat by: deficit, fiscal balance
同一指標、多個官方出口 / Same indicator, several official outlets
公屋「輪候幾年」——點解有兩個官方講法？
Why are there two official statements of how long the public rental housing queue takes?
房屋局同一份公布會列出兩個數：一般申請嘅綜合輪候時間，同撇除簡約公屋之後嘅數字。兩個都係官方數字，講「輪候時間」要講明用邊個口徑。輪候時間嘅官方定義（包括邊類申請唔適用）另見 fact 條目 prh-waiting-time-source。
One Housing Bureau release carries two figures: the composite waiting time for general applicants, and the figure with Light Public Housing excluded. Both are official, so the basis must be stated. The official definition, including which applicants fall outside it, is the separate fact entry prh-waiting-time-source.
5.1 年（2025年12月底） / 5.1 years (end-Dec 2025)
一般申請綜合輪候時間 / Composite waiting time for general applicants
公屋綜合輪候時間（一般申請） · asOf 2025-12 / Public housing comprehensive waiting time (general applications)
備註：房屋局一般申請綜合輪候時間。同一份公布嘅另一口徑（撇除簡約公屋）另見 fact 條目 prh-waiting-2025-excl-lph；輪候時間嘅官方定義見 prh-waiting-time-source。2026年6月底綜合輪候時間再降至4.8年。
出處：公屋綜合輪候時間年度分析 · 相關專題
5.6 年（2025年12月底） / 5.6 years (end-Dec 2025)
撇除簡約公屋之後嘅數字 / Figure with Light Public Housing excluded
公屋輪候時間（撇除簡約公屋） · asOf 2025-12 / Public housing waiting time (Light Public Housing excluded)
注意：唔可以同綜合輪候時間互換 / Note: Not interchangeable with the composite figure
備註：房屋局同一份公布嘅撇除簡約公屋口徑；一般申請綜合輪候時間另見 fact 條目 prh-waiting-2025。兩個口徑都係官方數字，唔可以互換，講「輪候時間」要講明用邊個。
出處：公屋綜合輪候時間年度分析 · 相關專題
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：公屋輪候、輪候時間、公屋要等幾耐 / Triggered in chat by: public housing waiting, prh waiting, waiting time
兩套官方口徑 / Two official bases
「香港一年捕到幾多魚」——係總漁獲量，抑或捕撈加養殖合計？
How much fish does Hong Kong take in a year — total catch, or catch plus culture output?
同一份官方年報嘅漁業概況會先後列出兩個範圍：一係「總漁獲量」（捕撈所得），一係「捕撈量同魚塘及箱網養殖量合計」（連養殖一齊計）。兩個係唔同範圍，唔可以相加，亦唔可以互相換算；講「漁獲」一定要指明係邊個口徑。
One official yearbook chapter prints two scopes one after the other: “total fish catch” (what the fleet lands), and “catch together with pond and cage culture output” (which includes fish farming). They are different scopes — not additive, not interchangeable — so state the basis whenever a catch figure is quoted.
約 92 250 公噸（2024 年） / about 92,250 tonnes (2024)
只計捕撈所得嘅總漁獲量 / Total catch from capture fisheries only
總漁獲量 · asOf 2024 / Total fish catch
備註：《香港年報2024》第十八章：二零二四年總漁獲量約92 250公噸，估計批發總值約24億元，其中供應本地食用嘅漁獲約35 000公噸。同一章另一行「捕撈量同養殖量合計」係94 340公噸，屬另一套口徑（見 /calibre#fishery-catch），兩個數唔可以相加或互換。
約 94 340 公噸（2024 年） / about 94,340 tonnes (2024)
捕撈量同魚塘及箱網養殖量合計 / Capture together with pond and cage culture output
捕撈量及養殖量合計 · asOf 2024 / Catch plus pond and cage culture output
注意：唔可以同「總漁獲量」相加，亦唔可以互相換算 / Note: Must not be added to, or converted into, the total catch figure
備註：《香港年報2024》第十八章漁業概況第一行：二零二四年捕撈量和魚塘及箱網養殖量合共約94 340公噸，總值25.2億元。同一段另一行嘅「總漁獲量」係92 250公噸，兩者係同一章兩個唔同口徑（見 /calibre#fishery-catch），唔可以相加或互相換算。
未齊全嘅部分：魚類養殖區同塘魚養殖屬另一組分項系統（另有獨立 fact），唔屬呢兩個口徑；官方亦冇公布兩者之外嘅第三個口徑。未核實之前唔會補數，只會寫入呢一段。 / Gap: Marine culture zones and pond fish farming are a separate set of sub-totals with their own facts and are not part of either basis here; no third official basis is published. Nothing is filled in until verified — it is recorded here instead.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：漁獲量、總漁獲、捕魚量、漁業產量 / Triggered in chat by: fish catch, total catch, fishery output
唔可以相加 / Not additive
「漁農業請幾多人」、「農業請幾多人」同「有幾多漁民」係同一條問題嗎？
Are “how many people farming and fishing employ”, “how many farms employ” and “how many fishermen” the same question?
官方文件各自公布一組數字：年報公布漁業同農業合計嘅直接僱用人數，漁護署兩版網頁分別公布漁業單獨嘅從業人數（只計捕魚作業嘅本地漁民）同農業單獨嘅直接僱用人數（連同農場數目）。四組數字時點同範圍都唔同：一份係兩業合計、兩份只計漁業（兩個唔同年份）、一份只計農業——既唔可以相加做「漁農業總人手」，亦唔可以相減去推任何一方。要講邊個就照抄邊個，並且標明時點。
Official documents each publish their own set: the yearbook gives one combined figure for fisheries and agriculture, while two AFCD pages separately give a fisheries-only figure (local fishermen engaged in fishing operations) and a farm-only direct-employment count (alongside a farm count). The four differ in both scope and reference point — two sectors combined, fisheries only (in two different years) and farming only — so they must not be added into a “total primary-sector workforce”, nor subtracted to infer any single sector. Quote whichever one is meant and state its reference point.
17 759 人（2024 年） / 17,759 (2024)
漁業同農業合計嘅直接僱用人數 / Persons directly employed in fisheries and agriculture combined
漁農業直接僱用人數（兩業合計） · asOf 2024 / Persons directly employed in agriculture and fisheries (combined)
備註：《香港年報2024》第十八章：本地漁農業合共直接僱用17 759人，生產總值34.48億元。呢個數字係「漁業＋農業」合計，官方冇分開兩業嘅僱用人數，所以唔可以當成單一行業數字。
約 10 180 名（2024 年） / about 10,180 (2024)
只計本地漁民 / Local fishermen only
本地漁民人數 · asOf 2024 / Local fishermen
注意：唔可以同漁農業合計相減去推農業僱員，亦唔可以同過港漁工配額相加 / Note: Must not be subtracted from the combined figure to infer farm employment, nor added to the Mainland deckhand quota
備註：《香港年報2024》第十八章：本地漁民約有10 180名；另有內地過港漁工配額4 110名。兩個數字係唔同身份嘅人力，唔可以相加當「漁業從業員總數」。
約 4 300 名農民及工人（漁護署 2026 年 5 月更新頁面所列「現時」數字） / about 4,300 farmers and workers (as stated currently on AFCD’s page updated in May 2026)
農業單獨嘅直接僱用人數（漁護署網頁，連同農場數目公布） / Farm-only direct employment as published on the AFCD page with the farm count
農業直接僱用人數 · asOf 2026 / Persons directly employed in agriculture
注意：唔可以同漁農業合計相加，亦唔可以相減去推漁業僱員；兩者年份亦唔同 / Note: Must not be added to the combined figure, nor subtracted from it to infer fisheries employment; the two also refer to different years
備註：漁農自然護理署〈香港的農業〉頁：「現時本港約有2 500個農場。它們直接僱用了約4 300個農民和工人。」呢個係農業單獨嘅「直接僱用」人數；同《香港年報2024》「漁農業合計直接僱用17 759人」係唔同時點、唔同範圍嘅兩組官方數字，唔可以相加或相減（見 /calibre#farm-fishery-employment）。
出處：漁農自然護理署《香港的農業》 · 相關專題
約 10 120 名（2025 年） / about 10,120 (2025)
漁業單獨嘅從業人數（漁護署捕撈漁業概況頁，只計從事捕魚作業嘅本地漁民） / Fisheries-only workforce as published on the AFCD capture-fisheries page (local fishermen engaged in fishing operations only)
捕魚作業本地漁民人數 · asOf 2025 / Local fishermen in fishing operations
注意：唔含養殖戶及魚類統營處職工，唔可以同農業數字相加，亦唔可以同漁農業合計相減 / Note: Excludes fish farmers and Fish Marketing Organisation staff; not to be added to the farm figure nor subtracted from the combined one
備註：漁農自然護理署〈香港捕撈漁業概況〉頁：「現時香港約有4 980艘漁船，並估計有約10 120名從事捕魚作業的本地漁民。」呢個係漁業單獨嘅從業人數，但只計捕魚作業（唔含養殖戶、魚類統營處職工等同業人手）；同《香港年報2024》第十八章嘅「本地漁民約10 180名」（2024年）係同一類數字嘅前後兩年，唔可以當成兩套口徑互相換算，亦唔可以同年報「漁農業合計直接僱用17 759人」（2024年）相加或相減（見 /calibre#farm-fishery-employment）。
未齊全嘅部分：未有嘅係「受僱僱員」同分業逐年嘅序列：年報只公布兩業合計（官方冇分開兩業嘅逐年僱員），漁護署嘅漁業同農業數字都係部門網頁嘅單一時點陳述，唔屬統計序列；內地過港漁工配額亦唔係從業人數。呢幾點留白，唔靠推算補。 / Gap: Still missing are an employee-only breakdown and any year-by-year series by sector: the yearbook publishes one combined figure without splitting the two sectors, the AFCD fisheries and farm counts are single-reference-point statements on departmental pages rather than a statistical series, and the Mainland deckhand quota is not a workforce count. These gaps stay open rather than being filled by inference.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：漁農業僱員、漁民人數、農業就業、漁業就業 / Triggered in chat by: farm employment, fishermen, fisheries employment
唔可以相加 / Not additive
「香港有幾多艘漁船」——係年底領牌嘅船隻數目，抑或漁護署講嘅作業漁船？
How many fishing vessels does Hong Kong have — vessels licensed at the year end, or the boats the AFCD counts as working the fisheries?
兩個官方數字量度唔同嘅事：年報公布嘅係年底領有漁船牌照嘅船隻數目（大大小小、唔同作業水域都算），漁護署捕撈漁業概況頁公布嘅係部門按「現時」作業情況描述嘅漁船數目。兩者時點同範圍都唔同，唔可以相加、相減或者互相換算。
The two official figures measure different things: the yearbook counts vessels holding fishing vessel licences at the year end (large and small, whatever waters they work), while the AFCD capture-fisheries page states how many vessels the department counts as working the fisheries at present. Differing in both scope and reference point, they must not be added, subtracted or converted.
6 125 艘（2024 年年底） / 6,125 (end of 2024)
年底領有漁船牌照嘅船隻數目 / Vessels holding fishing vessel licences at the end of the year
持漁船牌照船隻數目 · asOf 2024-12 / Vessels holding fishing vessel licences
備註：《香港年報2024》第十八章：截至二零二四年底，本港約有6 125艘船隻領有《商船（本地船隻）（證明書及牌照事宜）規例》下嘅漁船牌照，包括主要在南中國海作業嘅較大型漁船同在本港水域作業嘅較小型漁船。
約 4 980 艘（2025 年） / about 4,980 (2025)
漁護署「現時」描述嘅捕撈漁業作業漁船數目 / Vessels the AFCD states are currently operating in the capture fisheries
捕撈漁業作業漁船數目 · asOf 2025 / Vessels operating in the capture fisheries
注意：唔可以同持牌船隻數目相加或相減，亦唔可以當成同年數字 / Note: Must not be added to or subtracted from the licensed-vessel count, nor treated as the same year’s figure
備註：漁農自然護理署〈香港捕撈漁業概況〉頁：「現時香港約有4 980艘漁船，差不多全部是機動漁船」，其中舢舨及較小型非拖網漁船（刺網、延繩釣、圍網等）約佔本地漁船82%、拖網漁船及較大型非拖網漁船約佔18%。呢個係漁護署按「現時」作業情況描述嘅漁船數目，同《香港年報2024》第十八章「截至二零二四年底約6 125艘船隻領有漁船牌照」（發牌數目，年底時點）係兩套唔同嘅官方數字，唔可以相加、相減或互換（見 /calibre#fishing-vessel-counts）。
未齊全嘅部分：兩者之間嘅對照（例如領牌但唔出海作業嘅船隻）官方冇公布，本站唔會靠推算補一個差額。 / Gap: No official reconciliation exists between the two (for example vessels licensed but not fishing), and no difference is inferred here.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：漁船數目、有幾多艘漁船、持牌漁船、作業漁船 / Triggered in chat by: fishing vessels, how many fishing boats, licensed fishing vessels
唔可以相加 / Not additive
「香港有幾多個農場」——係全港各類農場總數，抑或持牌豬場同家禽農場嘅分項？
How many farms does Hong Kong have — the total across all farm types, or the licensed pig and poultry farms counted separately?
兩個官方數字係唔同層次：漁護署〈香港的農業〉頁講「現時本港約有…個農場」，係各類農場（菜場、果園、禽畜場等）合計嘅約數；政府一站通《香港便覽：農業及漁業》另列持牌豬場同家禽農場嘅數目，係禽畜分項。禽畜場已經包含喺前面嘅農場總數之內，所以兩者唔可以相加，亦唔可以相減去推其他農場類別。
The two official figures sit at different levels: the AFCD “Agriculture in Hong Kong” page gives an approximate total for farms of all kinds (vegetable, orchard and livestock farms together), while the GovHK factsheet gives licensed pig and poultry farms as a livestock sub-count. Livestock farms are already inside that total, so the two must not be added, nor subtracted to infer other farm types.
約 2 500 個（漁護署 2026 年 5 月更新頁面所列「現時」數字） / about 2,500 (as stated currently on AFCD’s page updated in May 2026)
全港各類農場合計（漁護署頁面「現時」約數） / All farm types combined (the AFCD page’s present-tense approximate figure)
農場數目 · asOf 2026 / Number of farms
備註：漁農自然護理署〈香港的農業〉頁（二零二六年五月更新）：「現時本港約有2 500個農場。」呢句係頁面用「現時」描述嘅狀況，唔係某一統計年嘅年報數字，引用時要連同時點一齊講。
出處：漁農自然護理署《香港的農業》 · 相關專題
豬場 43 個、家禽農場 29 個（2025 年） / 43 pig farms and 29 poultry farms (2025)
持牌豬場同家禽農場數目（禽畜分項） / Licensed pig farms and poultry farms (the livestock sub-count)
豬場及家禽農場數目 · asOf 2025 / Pig farms and poultry farms
注意：唔可以同全港農場總數相加，亦唔可以從總數相減去推其他類別農場 / Note: Not to be added to the all-farm total, nor subtracted from it to infer other farm types
備註：政府一站通《香港便覽：農業及漁業》（二零二五年十二月版）：「截至二零二五年，本地共有43個豬場及29個家禽農場。」呢兩個係持牌禽畜農場數目，同漁護署〈香港的農業〉頁「全港約有2 500個農場」（含菜場、果園等各類農場）係唔同範圍，唔可以相加或相減。同一頁另載「香港有兩個位於新界沙頭角及大埔的持牌奶場」。
出處：政府一站通《香港便覽：農業及漁業》 · 相關專題
未齊全嘅部分：禽畜以外各類農場（菜場、果園、花場等）嘅獨立官方數目冇公布，本站唔會靠減法補；同一份便覽另載兩個持牌奶場，未獨立立 fact，所以唔會喺呢組出現。 / Gap: No independent official counts exist for farm types other than livestock (vegetable, orchard, flower farms and so on), and none is inferred by subtraction; the same factsheet records two licensed dairy farms, which are not separate facts and therefore do not appear here.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：農場數目、豬場、家禽農場 / Triggered in chat by: number of farms, pig farms, poultry farms
兩套官方口徑 / Two official bases
「香港係唔係全球最大玩具出口地」——二〇〇四年官方新聞稿講「最大」，二〇〇七年同二〇一七年官方致辭講「第二大」，邊句先啱？
Is Hong Kong the world’s largest toy exporter? An official press release says largest, while two later official speeches say second largest — which applies?
三句都係官方講法，但年份同計算基礎唔同：二〇〇四年新聞稿冇講明係本地出口抑或連轉口計；二〇〇七年演講嘅出口總值明文「包括再次出口值」；二〇一七年演講就明言「若與轉口一併計算」先係全球第二大。所以三者唔可以互相取代——唔可以用二〇〇四年嘅「最大」講今日，亦唔可以把二〇〇七年嘅出口總值當成本地出口值。
All three are official statements, but their years and bases differ: the earliest press release does not say whether it counts domestic exports or re-exports too; the bureau speech states expressly that its export value “includes re-export values”; the later speech says Hong Kong is second largest only “counting re-exports”. The three therefore cannot stand in for one another — the earliest “largest” cannot be used for today, and the bureau’s export value cannot be read as a domestic export figure.
2004 年：全球最大的玩具出口地 / 2004: the world’s largest toy exporter
二〇〇四年投資推廣署新聞稿嘅講法（未講明計算基礎） / The InvestHK press-release wording, with no stated basis
「全球最大玩具出口地」官方講法（2004 年） · asOf 2004-03-13 / “World’s largest toy exporter” in an official statement (2004)
注意：唔可以當逐年統計，亦唔可以當今日嘅排名 / Note: Not a yearly statistic, and not today’s ranking
備註：投資推廣署2004年3月13日新聞稿：一間美國微型馬達製造商宣布在港設地區總部，新聞稿解釋「正因香港是全球最大的玩具出口地」；同一稿亦記「目前，全球約有一半的玩具出口來自香港」同「總值高達八十一億美元的香港玩具市場」。呢幾句係二〇〇四年新聞稿嘅講法，冇講明係本地出口抑或連轉口計，唔可以當逐年統計，亦唔可以同二〇〇七年「世界第二大」嘅講法互換（見 /calibre）。
出處：投資推廣署新聞稿：新一代微型馬達製造商選址香港作為亞太區策略性基地（「香港是全球最大的玩具出口地」） · 相關專題
超過 957 億港元（2007 年） / over HK$95.7 billion (2007)
二〇〇七年玩具出口總值，明文包括再次出口值 / The bureau’s toy export value, expressly including re-exports
玩具出口總值（2007 年，包括再次出口值） · asOf 2007 / Toy export value (2007, including re-exports)
注意：唔可以同只計本地出口嘅數字混用 / Note: Not to be mixed with domestic-export-only figures
備註：政制及內地事務局常任秘書長羅智光2008年10月9日出席香港玩具廠商會會員周年大會晚宴致辭：二〇〇七年香港玩具出口總值比二〇〇六年增長百分之二十三，總值超過九百五十七億港元，令香港保持世界第二大玩具出口地嘅地位。致辭明文寫「包括再次出口值」，即呢個數係連轉口嘅口徑，唔可以當成本地出口值，亦唔可以同只計本地出口嘅數字混用（見 /calibre）。官方冇公布逐年玩具出口值序列。
出處：政制及內地事務局常任秘書長出席香港玩具廠商會會員周年大會晚宴致辭全文（二〇〇七年玩具出口總值超過九百五十七億港元、世界第二大玩具出口地） · 相關專題
連轉口計：全球第二大（2017 年上半年） / with re-exports: second largest (first half of 2017)
二〇一七年上半年，連轉口計嘅排名 / The first half of the year in the later speech, ranking with re-exports counted in
玩具出口地世界排名（2017 年上半年，連轉口計） · asOf 2017-10-16 / World ranking of toy exports (first half of 2017, with re-exports)
注意：唔可以同二〇〇四年嘅「最大」互換，亦唔係全年數字 / Note: Not interchangeable with the earliest wording, and not a full-year figure
備註：政務司司長張建宗2017年10月16日出席美昌集團五十周年晚宴致辭：香港玩具出口於二〇一七年上半年錄得百分之二十六升幅，對歐盟出口更急增超過五成，「若與轉口一併計算，香港更是全球第二大玩具出口地」。呢句明言基礎係連轉口計，同二〇〇四年新聞稿嘅「全球最大」係唔同年份、唔同基礎嘅官方講法（見 /calibre）。
未齊全嘅部分：官方冇公布逐年玩具出口值或者逐年排名序列；二〇〇四年新聞稿亦冇交代計算基礎（本地出口抑或連轉口），所以兩邊對唔正，本站唔會據此推算任何一年嘅出口值。 / Gap: No year-by-year toy export value or ranking series is published, and the press release does not state its basis (domestic exports or re-exports included), so the two cannot be reconciled; no yearly export value is inferred here.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：玩具出口地排名、全球最大玩具出口地、世界第二大玩具出口地 / Triggered in chat by: toy export ranking, largest toy exporter, second largest toy exporter
同一指標、多個官方出口 / Same indicator, several official outlets
同一段時間嘅專營巴士數目，點解兩份運輸署刊物寫得唔同？
Why do two Transport Department publications give different franchised bus counts for the same period of the year?
兩份都係運輸署刊物，但統計時點唔同、亦係兩次唔同嘅車隊快照：《殘疾人士公共交通指南》（已領牌專營巴士，時點九月三十日）同《香港便覽—交通》（領牌專營巴士，時點十二月三十一日）。引用時一定要講明邊份、邊個時點，兩個數唔可以互相取代，亦唔可以相加。
Both are Transport Department publications, but their reference dates and fleet snapshots differ: the accessibility guide (licensed franchised buses, reference date the thirtieth of September) and the Hong Kong Fact Sheet on transport (licensed franchised buses, reference date the thirty-first of December). A citation must name the publication and the date; the two figures are neither interchangeable nor additive.
約 5,890 輛領牌巴士、約 760 條路線（四間專營巴士公司） / about 5,890 licensed buses and about 760 routes (four franchised bus companies)
《殘疾人士公共交通指南》第二章（二零二六年六月修訂本）所載車隊快照 / The fleet snapshot printed in chapter two of the accessibility guide (June revision of the publication year)
專營巴士車隊及路線（2024年9月底） · asOf 2024-09-30 / Franchised bus fleet and routes (end-Sep 2024)
注意：唔可以當年底數字，亦唔可以同便覽嘅數字混用 / Note: Not a year-end figure and not to be mixed with the fact sheet number
備註：運輸署《殘疾人士公共交通指南》第二章（2026 年 6 月修訂本）：截至 2024 年 9 月底，四間專營巴士公司共設 5,890 輛已領牌巴士、經營約 760 條巴士線。
5 間公司（6 項專營權）、約 5,870 輛 / 5 companies (6 franchises), about 5,870 buses
運輸署《香港便覽—交通》所載年底領牌專營巴士數目 / The year-end licensed franchised bus count in the Transport Department’s fact sheet
專營巴士公司及巴士數目 · asOf 2024 / Franchised bus companies and buses
注意：唔可以當成九月三十日嘅車隊，亦唔可以同指南嘅數字相加 / Note: Not the September fleet and not to be added to the guide’s figure
備註：運輸署《香港便覽－交通》：截至 2024 年 12 月底領牌專營巴士約 5,870 輛。
出處：運輸署：香港便覽—交通 · 相關專題
未齊全嘅部分：兩個數字之間嘅差額（車隊係唔係真係其間縮減）冇喺任何刊物頁面交代；兩份刊物亦冇逐字寫明係唔係包括非領牌／備用車輛。所以本站只可以講「時點唔同」，唔會推算一個「正確」數目，亦唔會話其中一個寫錯。 / Gap: Neither publication explains the difference between the two figures (whether the fleet really shrank in between), and neither states whether non-licensed or spare vehicles are included. The site therefore records only the different reference dates; it neither infers a “correct” count nor claims either figure is wrong.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：專營巴士數目、專營巴士幾多輛、專營巴士車隊 / Triggered in chat by: franchised bus fleet, number of franchised buses
同一指標、多個官方出口 / Same indicator, several official outlets
香港一年用幾多食水——點解官方同時講「全年總量」同「每日平均」？
How much fresh water does Hong Kong use in a year, and why do officials give both an annual total and a daily average?
兩個數字係同一批水量嘅兩種表述：一個係全年總量，一個係該總量除以日數得出嘅每日平均。所以唔可以相加（相加等於把同一批水計兩次），亦唔可以當成兩個獨立數字去講「增加」或「減少」；引用時要講明係年總量抑或日均，並保留原本嘅統計時點。
The two figures describe the same volume in two ways: an annual total and a daily average derived from it. Adding them counts the same water twice, and treating them as independent numbers invents a change that was never published. Say which basis you quote and keep its reference period.
1 075 百萬立方米 / 1,075 million cubic metres
政府統計處互動統計表轉載水務署數據嘅全年食水供應總量（按曆年點算） / Annual total fresh water supply for the calendar year, as carried in the C&SD interactive table from Water Supplies Department data
全年供水總量（2025年） · asOf 2025 / Total annual water supply (2025)
備註：政府統計處互動統計表 990-92072（水務署提供）：2025 年全年食水供應總量 1 075 百萬立方米。水務署另有「食水每日平均用量」fact（2.94 百萬立方米／日，年用量 1 074.59 百萬立方米）——兩者係同一水量嘅年總量同日均表述，唔可以相加，亦唔應該當成兩個獨立數字比較。
1 074.59 百萬立方米 / 1,074.59 million m³
水務署自行公布嘅全年食水用量 / Annual fresh water consumption as published by the department itself
全年食水用量（2025年） · asOf 2025-12-31 / Annual fresh water consumption (2025)
備註：水務署水資源數據：2025全年食水用量1,074.59百萬立方米（日均2.94），較2024年回升。
2.94 百萬立方米/日 / 2.94 million m³/day
同一批水量嘅每日平均用量 / The daily average of the same volume
食水每日平均用量（2025年） · asOf 2025 / Fresh water daily average consumption (2025)
注意：唔係額外一批水，同全年總量唔可以相加 / Note: Not additional water, and not to be added to the annual total
備註：2025年食水年用量1,074.59百萬立方米，日均2.94百萬立方米（按年+1.7%）；沖廁鹹水日均0.82百萬立方米。
未齊全嘅部分：上一個財政年度嘅全年用量屬另一個時段，本站另有 fact，唔屬於呢一組；兩個出口嘅數值只差進位，屬同一水量，所以唔可以當成「供應多過用量」嘅證據，亦冇逐年供應／用量序列可以對照。 / Gap: The preceding financial year’s consumption is a different period and a separate fact on this site, not part of this comparison. The two outlets differ only by rounding, so the difference must not be read as supply exceeding consumption; no year-by-year supply or consumption series is published here.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：食水供應、食水用量、供水總量、每日平均用量 / Triggered in chat by: water supply, water consumption, daily water use, fresh water
同一指標、多個官方出口 / Same indicator, several official outlets
食肆總收益有兩個同值條目，係唔係兩筆錢？
Two entries give the same restaurant receipts figure — are they two separate sums?
同一項官方統計經兩個出口公布：政府統計處嘅統計數字一覽，以及按年嘅餐飲服務業業務收益統計。兩者數值相同，只係分別擺喺經濟同飲食兩個專題，所以唔可以相加（相加會變成兩倍收益），亦唔可以當成兩個獨立指標去講增減。
One official statistic reaches the site through two outlets: the C&SD statistical digest and the annual business receipts statistics for food services. The two values agree and sit in the economy and food topics respectively; adding them doubles the receipts, and they are not two independent indicators.
1,096 億港元 / HK$109.6 billion
政府統計處統計數字一覽所載全年食肆總收益 / Annual restaurant receipts as carried in the C&SD statistical digest
食肆總收益價值 · asOf 2025 / Value of total restaurant receipts
出處：香港統計數字一覽 2026年版 · 相關專題
1,096 億港元 / HK$109.6 billion
政府統計處按年餐飲服務業業務收益統計所載同一指標 / The same indicator in the C&SD annual business receipts statistics for food services
食肆總收益（按年） · asOf 2025 / Total food-service receipts (annual)
備註：2025年食肆總收益約1,096億港元，屬政府統計處按年的餐飲服務業收益統計。
出處：香港便覽：零售業及食肆（政府統計處） · 相關專題
未齊全嘅部分：兩個出口嘅數值一致，但各自嘅分類明細（例如按食肆類型劃分）本站未登記，所以唔會由總數推算任何分項，亦唔會講邊個出口較準；時點同為該曆年。 / Gap: The two outlets agree, but their sub-breakdowns (by type of restaurant, for instance) are not registered here, so no component is derived from the total and neither outlet is presented as more accurate; both refer to the same calendar year.
問答提示觸發詞（Auto Chat 命中即喺答案提示呢組口徑）：食肆總收益、食肆收益、餐飲收益 / Triggered in chat by: restaurant receipts, food service receipts, catering receipts
Auto Chat
只從已核對條目作答，不發明數字。
