A table that continues onto the next page comes out as separate tables
You convert a listing that runs over several pages — and get one table for every page. The header row appears again at the top of each one, a "continued" caption or a page number sits between them, and in Excel the same table is spread over as many sheets as it had pages. Joining them back up by hand is what takes the time.
The cause is that a PDF is drawn one page at a time and records nothing about a table continuing. Its contents are instructions: "draw this glyph at these coordinates", "draw a rule here". Nothing in the file says that the rows on this page carry on from the rows on the last. The header is printed again on every page and a page number or a "continued" caption sits in between, so read mechanically, one table becomes as many tables as it had pages.
So the continuation has to be read back out of the run of pages. PDFIntact compares the last table on one page with the first table on the next, and where the only things between them can be skipped — a page number, a "continued" caption — it joins them into a single table with the header row kept just once. A joined table also carries the range of pages it spanned.
Try it now
Check it on your own PDF
The check runs entirely inside your browser. Your file is never uploaded, and no account is needed. See how many pages, and how many tables and figures, can be extracted before you pay.
Example
3つの表が、1つの表になります
3ページに分かれて出ていた表が、どう1つになるかを示した図です。実際の値ではなく、行の並び方だけを表しています。
そのまま変換したとき（ページごとの表）
［1ページ目の表］ 見出し行 行 A 行 B ［2ページ目の表］（続き） 見出し行 ← 2度目 行 C 行 D ［3ページ目の表］ 見出し行 ← 3度目 行 E
PDFIntact を通したとき（1つの表）
［1つの表］跨いだページ 1〜3 見出し行 行 A 行 B 行 C 行 D 行 E
3ページに跨る統計表で実際に確かめています。 3ページに分かれていた表が1つの表になり、見出し行は1行だけ、続く165行のデータは原本と同じ順序で並びました。
繋がない表もあります
同じ行のままページごとに列（項目）が変わる型の表は繋ぎません。 行として繋ぐと壊れるからです。その型の検体でも、3ページとも独立した表のままで、 誤って繋ぐことはありませんでした。 続きを示す語も「（続き）」のように括弧の中か行の端にある形だけを見ます。 「手続き」「続柄」のような語を含む本文で、無関係な表を繋いでしまわないためです。
形式ごとに、どう入るか
繋いだ表は1つの表として入りますが、跨いだページの範囲が残る形式は限られます。
Excel（.xlsx）
1つの表が1枚のシート。跨いでいた表も1枚に収まり、見出し行は1行です。シート名は表が始まったページの番号から付きます。罫線と結合セルは再現します。
Word（.docx）
文書順のまま、表が始まったページの位置に1つの表として入ります。
Markdown（.md）
同じ位置に、パイプ記法の1つの表として置きます。
JSON
繋いだ表にだけ、跨いでいたページの範囲（先頭と末尾）が入ります。繋いでいない表には付きません。
吸い上げられたページに残るもの
繋いだ行は先頭のページの表にまとまるので、後ろのページには表が残りません。 残るのは「（続き）」のキャプションだけで、それは本文のブロックとして出ます。 跨いでいたページは表の単位では辿れますが、行の単位では分かりません。 また、表として繋がることと、中の文字が正しく読めていることは別の話です。 ブロックごとの確度の帯を見て、結果は必ず元のPDFと照合してください。
Frequently asked questions
- Will a table that continues onto the next page come out as one table?
- That is what it is for. The last table on one page and the first table on the next are compared, and where the only things between them can be skipped — a page number, a caption reading "continued" — the rows are joined into a single table. In our testing a listing spread over three pages came out as one table of 165 data rows, in the original order. Where the two do not line up they are left as separate tables rather than forced together.
- Does the header row get repeated at the join?
- No. The header printed again at the top of each page is not carried into the joined table; it is kept once, at the top. In the three-page listing we tested, the joined table had a single header row, with the data rows below it in the original order.
- What about a table whose columns continue on the next page?
- Those are not joined. A table that keeps the same rows and puts different columns on each page would be destroyed by joining it as rows, so it stays as one table per page — and we checked on a specimen of exactly that shape that it is not joined by mistake. It comes out page by page, with the caption as the only sign that the pages belong together.
- Can I tell which pages the table spanned?
- For the table as a whole, yes: a joined table carries the first and the last page it spanned in the JSON output. Which page an individual row came from is not recorded. Note too that the pages taken up into the join no longer carry a table of their own — what stays on them is the "continued" caption, as a block of text.
- How does it come out in Excel?
- One table is one sheet, so a table joined across three pages arrives as a single sheet with one header row, named after the page where the table begins. No sheet is created for the pages taken up into it. Rules and merged cells are reproduced. The range of pages the table spanned is not written into the Excel file; it is in the JSON output.
Check any PDF against all symptoms／ What each format contains