You convert a cleanly formatted PDF table to Excel, open the output, and find that what was two separate columns of numbers in the PDF has become a single column with values like "1,2503,780" in each cell. Two numeric values from adjacent columns have been concatenated into one text string, and the data is now useless for calculation. This is the most common complaint about PDF to Excel conversion, and it happens because the converter misjudged where one column ends and the next begins.
Column boundary detection in PDF tables is a fundamentally hard problem because PDFs do not contain table structure. A PDF table is just text positioned at specific coordinates, with horizontal and vertical lines drawn near the text. The converter must infer the column boundaries from the spacing between text blocks, from the alignment of text across rows, or from the position of the drawn lines. When numeric columns are narrow and the numbers in adjacent columns are close together, the converter may merge them into a single wider column, reading both numbers as one text string. The converter sees a continuous horizontal run of numbers and cannot tell where the column break was in the original layout.
Why Adjacent Numeric Columns Are Especially Vulnerable to Merging
Numeric columns in PDF tables tend to be narrow because numbers are short. A column of currency values might contain entries like "1,250.00" or "42.50," rarely exceeding 12 characters. Two narrow numeric columns side by side, such as "Unit Price" and "Extended Price," often have just a few points of white space between them. If the converter's column-splitting algorithm uses a minimum gap threshold that is larger than the actual gap between the columns, it sees the two columns as one and concatenates their values.
The problem is compounded by right-aligned numbers. In a well-formatted PDF table, numeric columns are right-aligned, so the numbers in each row end at the same horizontal position. The space between the end of a right-aligned number in column A and the beginning of a left-aligned number in column B can be as little as a few points. Visual inspection of the PDF shows a clear gap, but the converter's algorithm may require a larger gap to confidently identify a column boundary, treating the narrow space as normal word spacing rather than a column separator.
A related issue occurs with negative numbers and parentheses notation. A value displayed as "(1,250.00)" includes both a parenthesis and a comma, which a converter may interpret as multiple separate tokens. The opening parenthesis gets associated with the previous column, the numeric portion gets placed in the current column, and the closing parenthesis gets dropped or attached to the next column. The result is a cell containing partial data that requires extensive manual cleanup. Tables using European number formatting, where the comma is a decimal separator and the period is a thousands separator, further confuse converters that assume US number formatting.
How to Identify Problem Tables Before Conversion
Before converting, open the PDF and zoom in on the table area. Look for columns where the space between the rightmost character of one column and the leftmost character of the next column is visually small, less than approximately one character width. These are the columns most likely to merge during conversion. If you see tables with these tight column spacings, plan for manual correction after conversion or consider an alternative extraction method.
Also check for tables that use leader dots, rows of periods connecting a left-aligned item name to a right-aligned price. Leader dots sit in the space between columns and are frequently misread as data rather than as visual formatting. A converter that treats leader dots as content will produce a column containing nothing but periods, or will split the table incorrectly because it cannot distinguish the leader dots from the data. Tables with merged header cells spanning multiple columns are another trouble spot. The converter sees a wide text block in the header row and may not correctly map it to the narrower data columns below.
For critical data where accuracy matters, Extract PDF Data extraction using specialized table recognition software is more reliable than general-purpose PDF-to-Excel conversion. Data extraction software designed specifically for tables uses multiple signals, text position, font metrics, line position, and alignment consistency across rows, to build a more accurate column model. The extra cost of specialized extraction software is justified when the alternative is hours of manual Excel cleanup for each converted table.
Manual Correction Strategies for Merged Columns
When merged columns are unavoidable, Excel's Text to Columns feature is the fastest correction tool. Select the merged column, open Data, Text to Columns, and choose Delimited. If the merged values are consistently separated by a space, choose Space as the delimiter. If they are separated by a variable number of spaces, choose Fixed Width and manually place the column break line between the two values. Excel previews the split before applying it, so you can adjust the break position until the split is correct for all rows.
For values that were merged without any delimiter, such as "12503780" where the split should be between "1250" and "3780," Text to Columns cannot help because there is no delimiter to split on. You need to know the expected width of each column. If the first column always contains four digits and the second always contains four digits, use Excel's LEFT and RIGHT functions with the known character counts to extract the values into separate columns. This approach works only when column widths are fixed and consistent, which is common in tables exported from database systems with fixed field widths.
WukongPDF converts PDF tables to Excel with column detection that analyzes text alignment, spacing, and line position to identify column boundaries even in tightly spaced numeric tables. The conversion preserves numeric formatting so that extracted values are immediately usable for calculation without manual cleanup. For complex tables that resist automated column detection, consider using an OCR-based approach as an alternative path. Capture a screenshot of the PDF table, run it through an OCR tool with table recognition capabilities, and export the recognized table to Excel. Modern OCR engines that include table structure detection are often more reliable at column separation than traditional PDF-to-Excel converters because they analyze the visual layout directly.