Extract Data from Table in PDF with Java

Extract tables from PDF

Use TableAbsorber to find tables on each page and iterate through rows, cells, text fragments, and text segments.

1. Open the source PDF in a Document instance.
2. Iterate through the document Page objects because tables are detected page by page.
3. Create a TableAbsorber for each page and call visit(page) to populate the detected table list.
4. Iterate through the detected AbsorbedTable, AbsorbedRow, AbsorbedCell, TextFragment, and TextSegment objects.
5. Build the extracted row text from the fragment content and print the table data.

public static void extractTablesFromPdf(Path inputFile) {
    try (Document document = new Document(inputFile.toString())) {
        for (Page page : document.getPages()) {
            TableAbsorber absorber = new TableAbsorber();
            absorber.visit(page);

            for (AbsorbedTable table : absorber.getTableList()) {
                System.out.println("Table");
                for (AbsorbedRow row : table.getRowList()) {
                    StringBuilder rowText = new StringBuilder();
                    for (AbsorbedCell cell : row.getCellList()) {
                        if (rowText.length() > 0) {
                            rowText.append("|");
                        }
                        StringBuilder cellText = new StringBuilder();
                        for (TextFragment fragment : cell.getTextFragments()) {
                            StringBuilder fragmentText = new StringBuilder();
                            for (TextSegment segment : fragment.getSegments()) {
                                fragmentText.append(segment.getText());
                            }
                            if (cellText.length() > 0) {
                                cellText.append("|");
                            }
                            cellText.append(fragmentText);
                        }
                        rowText.append(cellText);
                    }
                    System.out.println(rowText);
                }
            }
        }
    }
}

Extract a table from a specific marked area

This example finds a square annotation, compares its rectangle to each detected table, and outputs only tables inside the marked region.

1. Open the source PDF in a Document instance.
2. Get the target Page and locate the square Annotation that marks the extraction region.
3. Create a TableAbsorber and call visit(page) to detect tables on that page.
4. Compare each detected AbsorbedTable Rectangle with the annotation rectangle bounds.
5. Iterate through the matching AbsorbedRow and AbsorbedCell objects and reconstruct the row text.
6. Print the table data for the marked region only.

public static void extractTableFromSpecificArea(Path inputFile) {
    try (Document document = new Document(inputFile.toString())) {
        Page page = document.getPages().get_Item(1);

        Annotation squareAnnotation = null;
        for (Annotation annotation : page.getAnnotations()) {
            if (annotation.getAnnotationType() == AnnotationType.Square) {
                squareAnnotation = annotation;
                break;
            }
        }

        if (squareAnnotation == null) {
            System.out.println("No square annotation found.");
            return;
        }

        TableAbsorber absorber = new TableAbsorber();
        absorber.visit(page);

        for (AbsorbedTable table : absorber.getTableList()) {
            Rectangle tableRect = table.getRectangle();
            Rectangle annotationRect = squareAnnotation.getRect();

            boolean isInRegion = annotationRect.getLLX() < tableRect.getLLX()
                    && annotationRect.getLLY() < tableRect.getLLY()
                    && annotationRect.getURX() > tableRect.getURX()
                    && annotationRect.getURY() > tableRect.getURY();

            if (isInRegion) {
                for (AbsorbedRow row : table.getRowList()) {
                    StringBuilder rowText = new StringBuilder();
                    for (AbsorbedCell cell : row.getCellList()) {
                        if (rowText.length() > 0) {
                            rowText.append("|");
                        }
                        StringBuilder cellText = new StringBuilder();
                        for (TextFragment fragment : cell.getTextFragments()) {
                            StringBuilder fragmentText = new StringBuilder();
                            for (TextSegment segment : fragment.getSegments()) {
                                fragmentText.append(segment.getText());
                            }
                            if (cellText.length() > 0) {
                                cellText.append("|");
                            }
                            cellText.append(fragmentText);
                        }
                        rowText.append(cellText);
                    }
                    System.out.println(rowText);
                }
            }
        }
    }
}

Export tables to Excel

1. Open the source PDF in a Document instance.
2. Create ExcelSaveOptions for the export.
3. Set the Excel output format to XLSX so detected table layout is written as an Excel workbook.
4. Call document.save(outputFile.toString(), excelSave) to export the document in Excel format.

public static void exportTablesToExcel(Path inputFile, Path outputFile) {
    try (Document document = new Document(inputFile.toString())) {
        ExcelSaveOptions excelSave = new ExcelSaveOptions();
        excelSave.setFormat(ExcelSaveOptions.ExcelFormat.XLSX);
        document.save(outputFile.toString(), excelSave);
    }
}