Rent roll PDF to CSV, with the totals proved

A rent roll with totals proved is a spreadsheet-ready copy of a rent roll PDF in which the extracted unit rows add up to the totals printed on the original page. RentRoll Engine produces one: it reads the unit table, writes a fixed-column CSV, then sums monthly rent and square feet and compares both with the Total line on the report.

That comparison is the product. Getting text out of a PDF is easy. A dropped row, or a 3 read as an 8, still gives a tidy spreadsheet, and that spreadsheet then feeds an underwriting model, a debt service coverage ratio and a credit memo. When the rows tie to the page, you can say so. When they do not, you see the variance before you build on it.

What it proves is narrow, and the page says so. It shows that the CSV matches the PDF. It does not show that the PDF matches the leases, the general ledger or the bank.

1. Upload the rent roll PDF

   Drop the file from Yardi, RealPage, AppFolio, MRI, Buildium or any system that prints a unit table. Text PDFs and scans both work, up to the page limit of your plan.

2. The tables are rebuilt

   The engine reads the page tables, using OCR when the PDF is a scan. A rent roll that runs over several pages is joined into one roll: repeated headings and continuation pages are recognised.

3. Headings are mapped to fixed columns

   Unit or Suite, Tenant or Resident, SF or RSF, Rent, Lease From and To, Deposit. Market rent, rent per square foot and unit type are not rent, so they are left out and listed so you can see what was ignored.

4. The rows are summed against the page

   Monthly rent and square feet are added up and compared with the Total line the report prints. The banner is green when they agree and amber, with the variance, when they do not.

5. Download the CSV

   Ten columns in a fixed order, ISO dates, plain decimals, and a flag on every row that needed a judgement. JSON is included on Pro.

These are the two samples above, produced by the same code that runs on your file. Both properties are fictional. The first rent roll ties. The second is an office roll with an annual rent column, and it does not.

In the office roll the five occupied rows sum to $27,680.00 a month, and the page total works out to $27,830.00. The difference is $150 a month, or $1,800 a year. The tool cannot say whether a row is missing, a figure was keyed wrongly or the printed total is stale. It does say, before the number reaches a model, that the rows and the page disagree and by how much. If that $1,800 flowed through to NOI, a 7% cap rate would put it at about $25,700 of price.

One row, before and after

Printed on the page
| 100   | Harbor Dental Group        | 2,400  | Jan 1, 2022 | Dec 31, 2027     | $62,400.00  | $26.00           | $10,400.00       | Two 5-yr options, 3% annual bumps |

Written to the CSV
100,Harbor Dental Group,occupied,2400,5200.00,2022-01-01,2027-12-31,10400.00,"Two 5-yr options, 3% annual bumps",rent_annual_to_monthly

The page prints annual rent, so $62,400.00 becomes 5200.00 a month and the row is flagged rent_annual_to_monthly. The rate per square foot is not rent, so it is left out.

Every file has the same ten columns in the same order, so a formula or Power Query step written once keeps working.

Rules the file follows

• UTF-8, comma separated, CRLF line endings, fields quoted only when they need to be (RFC 4180). No byte-order mark.
• Money has two decimals and no symbol. Negative amounts in parentheses, (1,234.56), or with a trailing minus are written as -1234.56.
• European number formats such as 1.234,56 are recognised when the rest of the page shows the same convention.
• Dates are ISO. When 04/05/2025 could be April 5 or May 4 and nothing else on the page settles it, the row is flagged date_ambiguous instead of guessed silently.
• A text cell that starts with =, +, - or @ is written with a leading apostrophe, so a spreadsheet cannot run it as a formula.

Want to fill the same layout in by hand? Download the rent roll template.

The reader works from the meaning of the column headings, not from the name of the system that printed them. These are the systems rent rolls most often come from.

• Yardi. Resident codes, market versus actual rent, charge lines under each unit.
• RealPage. Floor plan codes, status columns, summary tables after the roll.
• AppFolio. Status values, delinquency columns, rolls with no area column.
• MRI and Buildium. Read by the same table reader. There is no dedicated guide yet.
• Rent rolls built in Excel and printed to PDF, from owners and brokers.

Yardi, RealPage, AppFolio, MRI and Buildium are trademarks of their owners. RentRoll Engine is independent of all of them.