K-1 data extraction means reading every amount and code from a Schedule K-1, statements included, into rows a preparer can sort and total. Identify the version first (1065, 1120-S or 1041), capture one row per box and code with its source page, and treat any box that points to a statement as unfinished until that statement is read.
K-1 data extraction across the 1065, 1120-S and 1041 versions
K-1 data extraction starts by telling three forms apart. A Schedule K-1 from a partnership (Form 1065), an S corporation (Form 1120-S) and an estate or trust (Form 1041) share a name, but their box numbers differ. On the 2025 forms, box 5 is interest on a partnership K-1, while box 5a is ordinary dividends on an S corporation K-1.
The quickest test is the title line. On the 2025 forms, the Schedule K-1 (Form 1065) reads "Partner's Share of Income, Deductions, Credits, etc.", the Schedule K-1 (Form 1120-S) reads "Shareholder's Share", and the Schedule K-1 (Form 1041) reads "Beneficiary's Share". The form number in the top left corner confirms it.
The page structure differs too. Each version has the same three parts: Part I about the entity, Part II about the recipient, and Part III with the current-year amounts. What sits inside each part changes:
- 1065, 2025. Part II runs from item E to item N, including item J (profit, loss and capital percentages), item K1 (liabilities) and item L (the capital account analysis). Part III runs from box 1 to box 23.
- 1120-S, 2025. Part II runs from item E to item I, including the allocation percentage, share counts and loans from the shareholder. Part III runs from box 1 to box 19.
- 1041, 2025. Part I includes item E, checked when this is the final Form 1041 for the estate or trust. Part III runs from box 1 to box 14.
Record the form year alongside the version. Each year's K-1 comes with its own instructions and code lists, so a template keyed to 2025 codes needs checking against the 2026 instructions before it is reused.
What each box reports, version by version, is set out in Schedule K-1 boxes explained, starting with the boxes on the partnership Form 1065 K-1. Extraction only needs to capture the box, the code and the amount faithfully; interpreting them is the next step.
K-1 OCR on the face page versus the supplemental statements
K-1 OCR on the face page is the easy part because the face page is a fixed IRS layout: the same boxes in the same places on every 2025 K-1 of one version. Supplemental statements are the hard part. They are pages the entity writes itself, in its own format, and they often hold most of the detail.
The face page leads to the statements in plain terms. On the 2025 Schedule K-1 (Form 1065) and Schedule K-1 (Form 1120-S), the last two boxes carry an asterisk with the note "See attached statement for additional information." The 2025 Schedule K-1 (Form 1041) adds: "A statement must be attached showing the beneficiary's share of income and directly apportioned deductions from each business, rental real estate, and other rental activity."
The Partner's Instructions for Schedule K-1 (Form 1065) (2025) explain one reason: "If the partnership had more than one activity, it'll attach a statement to your Schedule K-1." For a partnership with a rental and an operating business, that statement is where amounts are broken out by activity.
Read the face page and the statements as one document. A face page read alone gives totals. The statements give the breakdown the return may need: by activity, by code, by state. Picture an example partnership packet of 14 pages: two pages of IRS form and 12 of statements. The statement pages deserve the same care as the face.
Practical steps for each K-1 package:
- Count the pages and note which are the IRS form and which are statements.
- Read the face page into box-and-code rows.
- For every box with a code or a statement reference, find the matching statement page.
- Add the statement detail as its own rows, linked to the face-page row.
- Compare the statement rows with the face-page figure they explain, and note any gap.
The last step catches two problems at once. If statement rows add up to more than the face-page box, a page was read twice. If they add up to less, a page is missing or was skipped. Neither shows up when the face page is read on its own.
Corrected K-1s need the same treatment as corrected 1099s. The 2025 K-1s each carry an "Amended K-1" check box at the top. When an amended package arrives, mark the earlier rows for that entity and year as superseded rather than deleting them, so the change stays visible to the reviewer.
How do you capture K-1 footnotes and attachments?
K-1 footnotes and attachments are the pages after the IRS form: the entity's supplemental statements, any state schedules, Section 199A detail and, when checked, a Schedule K-3. Capture them as rows in the same sheet as the face page, each tagged with the box and code it explains and the page it came from.
Section 199A information shows how much lives off the face page. The 2025 partner's instructions say of box 20, code Z: "The partnership will provide the information you need to figure your deduction." The Shareholder's Instructions for Schedule K-1 (Form 1120-S) (2025) say box 17, code V includes qualified business income items, W-2 wages, unadjusted basis of qualified property and Section 199A dividends. That is several figures behind one code, so each needs its own row.
The Instructions for Schedule K-1 (Form 1041) (2025) put the beneficiary's Section 199A information under box 14, code I.
Schedule K-3 is flagged on the face page itself. The 2025 partnership K-1 has box 16, "Schedule K-3 is attached if checked", and the S corporation K-1 has the same check box as box 14. When the box is checked, the K-3 pages belong to the same package and should be counted as part of it.
State schedules vary by entity and state. Instead of assuming one exists, add a line to the request: ask whether a state K-1 or state schedule came with the federal one, and list each by state when it did.
Footnotes written as text need a home too. A statement line such as "Box 13, code ZZ: other deductions, see detail" followed by a short list should become one row per item, with the text kept in a notes column so the reviewer sees the entity's own wording.
The interest and dividend boxes on a K-1 look like 1099 amounts but come from a different form. Keep them on the K-1 sheet. For how interest and dividends differ when they arrive on their own forms, see 1099-INT vs 1099-DIV.
A K-1 to Excel layout for clients with many K-1s
A K-1 to Excel layout for clients with many K-1s works best as one long table with one row per box and code, never one column per box. A column-per-box sheet breaks as soon as two K-1s use the same box with different codes, or a statement adds five lines under one code.
Each row answers the same questions: which entity, which version and year, which box, which code, what amount, and where on the PDF it came from. Totals by box and code across every K-1 then come from a pivot table, and a reviewer can filter to one entity in seconds.
K-1 to Excel template, filled with example rows
The entities, EINs and amounts below are fictional examples for a fictional client, Priya Sandoval.
| Entity | EIN | K-1 version | Form year | Box | Code | Amount | Source page | Notes |
|---|---|---|---|---|---|---|---|---|
| Canopy Partners LP | XX-XXX4821 | 1065 | 2025 | 1 | 18,240.00 | canopy.pdf p1 | Two activities, see p4 | |
| Canopy Partners LP | XX-XXX4821 | 1065 | 2025 | 1 | 11,900.00 | canopy.pdf p4 | Activity 1, operating | |
| Canopy Partners LP | XX-XXX4821 | 1065 | 2025 | 1 | 6,340.00 | canopy.pdf p4 | Activity 2, operating | |
| Canopy Partners LP | XX-XXX4821 | 1065 | 2025 | 20 | Z | STMT | canopy.pdf p1 | Detail on p6 |
| Canopy Partners LP | XX-XXX4821 | 1065 | 2025 | 20 | Z | 17,480.00 | canopy.pdf p6 | QBI, activity 1 and 2 |
| Bellwether Dental Inc | XX-XXX7310 | 1120-S | 2025 | 17 | V | 42,115.00 | bellwether.pdf p5 | QBI |
| Hollis Family Trust | XX-XXX0912 | 1041 | 2025 | 14 | I | STMT | hollis.pdf p1 | Detail on p3 |
Two habits make the sheet trustworthy. Mark face-page rows and statement rows so totals do not double count; in the example, the two activity rows add to the face-page box 1 figure and should replace it in totals, not sit beside it. And keep the source page on every row, because the first question in review is always "where did this come from?"
A client with eight K-1s is where this layout pays off. Schema CPA reads K-1s alongside the 1099s, W-2s and statements in the same client folder and exports to Excel or CSV, so the rows start from the PDF rather than from typing. The client tracker demo shows each client's documents marked as received, missing or duplicate.
Why does K-1 data extraction miss amounts marked STMT?
K-1 data extraction misses amounts marked STMT when the face page is read alone and a box that says STMT, "see attached" or "see statement" is captured as blank or zero. STMT is not an IRS term; it is the shorthand some entities print in a box to point at the supplemental statement. The amount exists, just on another page.
The mistake is easy to make. A face page with "STMT" in box 20 looks complete at a glance: every box is filled with something. An OCR pass that expects numbers may skip the text, and a person typing quickly may treat it as a note. Either way the return loses the figures behind the code.
The form gives the warning in its own words. The 2025 partnership and S corporation K-1s mark boxes 22 and 23, and boxes 18 and 19, with "See attached statement for additional information." Other items point to statements too: item M on the 2025 partnership K-1 asks whether the partner contributed property with a built-in gain or loss and says "If 'Yes,' attach statement."
Three rules prevent the miss:
- Never write zero for text. If a box holds STMT or similar, the row's amount stays as STMT until the statement row replaces it.
- Check for open STMT rows before review. Filter the sheet for STMT. Any row still showing it has an unread page behind it.
- Treat a missing statement as a missing document. If the PDF has no page matching the reference, the K-1 package is incomplete and the entity owes the rest.
A K-1 that is incomplete or not yet issued is a timing problem as well as a data one. When to wait, extend or estimate is covered in when a K-1 arrives late.