PDF to XML
The PDF to XML page shows the XML a PDF produces, to start or adjust the mapping of a PDF-source document. It converts a text PDF from any system — JD Edwards, SAP, NetSuite, a custom ERP or any invoicing tool — or a JD Edwards tagged PDF, with or without a document's settings, and saves to the document the settings tried on the page. Nothing is stored by a conversion: the page is a working tool.
Access: Processing → PDF to XML.
At a glance
Converting a PDF
| Control | Description |
|---|---|
| Choose a PDF | Picks the PDF on the computer and converts it at once. |
| Converter | Text PDF (a PDF whose text can be selected, from any generator) or JD Edwards PDF (a BI Publisher / JD Edwards print carrying its section and field tags). |
| Document | Default settings, or a document whose settings are loaded and tried. Picking a document also sets the converter: Text PDF for a PDF source, JD Edwards PDF otherwise, with the manifest of the watched folder that uses the document. |
| New document | Creates a document for the converter in use — Name (letters, digits, - and _) and Description. Its UBL XSLT points to <name>.xsl, to create in the XSL editor if needed; the settings tried on the page stay pending, ready for Save to document. |
| Convert | Converts again with the settings currently on the page, saved or not. |
| Download XML | Saves the XML as <pdf name>.xml. |
The XML appears on the right, read-only. A scanned PDF (images without text) is refused with no text found … (scanned or empty PDF?).
Text PDF
Invoice lines
Table start, Header rows and Table end set the line table. The easiest way is to click the rows in the left pane, with Invoice lines selected:
- Click the first row of the table header;
- Click the last row of the table header;
- Click the first row after the lines.
The page writes the three settings and converts again: the header rows are framed in indigo, the lines in green, the end row in red, and the count shows N line(s) found. Start over clears the clicks.
VAT summary
With VAT summary selected, the same clicks set VAT table start, Columns and VAT table end: Click the header row of the VAT summary, Click its last column (one button per header cell — useful when another block is printed beside the summary), Click the first row after the rates. The rates become <Recap_TVA>/<Ligne_TVA>, and the count shows N VAT rate(s) found.
Document key
| Control | Description |
|---|---|
| Invoice number | The element holding the invoice number. A number such as 36834.FU (number then letters) or FU-36834 is split into Number and Type, both read from the label printed before the value. |
| Document type | Shown when the number cannot be split: the fixed document type (e.g. RI). |
| Several invoices | For a PDF holding several invoices: the rule that gives one document per invoice number (One invoice per PDF when empty). Split on this number builds it from the label of the invoice number; the count then shows N invoice(s). |
| Find company by | Not searched, VAT number, SIRET or SIREN, then the element holding the identifier: the page shows the company found among the supplier companies (→ 00001), or why none matches. |
| Company | Fixed company code, when the company is not searched. |
Save to document
Save to document writes the settings to the document picked: table start, header rows and table end, VAT summary, invoice split, the number and type extraction rules, the company (or its lookup). A document without a burst key gets Document, one per invoice. The same settings can be edited afterwards in the PDF extraction group of the document.
JD Edwards PDF
| Control | Description |
|---|---|
| Manifest | Server path of the manifest that names the tags; empty = data dictionary names. The folder button browses the server. |
| Starter manifest | Downloads a manifest built from the PDF, to complete. |
| Invoice key | For a report that does not follow the standard invoice layout: the section:field holding the invoice number, shown with the number of distinct values. Gives one <Invoice> per value (burst key Invoice). None (standard sections) keeps the standard layout. |
| Burst key, Document ID, Document type, Company | Picked among the elements of the XML; Find company by replaces the fixed company as for a text PDF. The count shows N invoice(s). |
| Save to document | Writes the invoice key, burst key, document ID, type and company (or its lookup) to the document. |
A PDF without JD Edwards tags shows No JD Edwards tags in this PDF: it needs PDF compression turned off in the JDE INI.
A JD Edwards PDF can also be processed directly from Process Document on an XML-source document: it is converted with the document's invoice key and the manifest of its watched folder.
Access
The page follows the Process Document page grant on the Roles page: a user who can process documents sees it.
Tips & best practices
- Start from a real PDF of each layout. One sample per layout is enough to set the table, the VAT summary and the key.
- Click rather than type the anchors. The page writes the shortest rule that finds the row, numbers turned into
\d+, so the next PDF of the same layout matches too. - Check the counts before saving. Lines found, VAT rates found and invoices must match the PDF; a red count means an anchor is wrong.
- Search the company when one PDF layout serves several companies. The lookup by SIRET, SIREN or VAT number avoids one document per company; an identifier missing from the supplier companies stops the invoice.
- Use the AI Auto-Map for the values the converter merges. Its extraction rules are saved on the document and show up here on the next conversion.