How to Upload Data into SAP from Excel

PostNow.ai uploads data into SAP from Excel: it maps the sheet to the SAP object, validates every row against the live system and posts through standard BAPIs under your own SAP login, with the result written beside each row. If the data already sits in a spreadsheet, keying it into SAP a screen at a time is the slowest and least reliable way to get it there. This guide is the procedure: how to upload data into SAP from Excel in eight steps that work for every object and every load method, with the decisions that matter at each step and the errors that appear when a step is skipped. It uses one worked example throughout, a batch of 40 journal entry lines, so the steps are concrete rather than abstract. Whether you call it an upload or an SAP Excel import, and whether the search was "SAP import data from Excel" or "SAP import Excel", the procedure is the same; for the tools that automate it, see Excel to SAP automation.

It is written for the person doing the load: a finance user with a month end batch, a master data lead with 3,000 materials to extend, a consultant with a cutover file. If you want the comparison of load methods rather than the procedure, that is a separate guide, SAP data upload methods, and the SAP mass upload guide covers volume, batching and governance in depth. This page stays on the procedure.

What an Excel to SAP upload actually does

PostNow.ai is the add-in that sits between Excel and SAP and does this work. An Excel upload in SAP is not Excel talking to SAP. Excel holds rows; SAP holds business objects with rules. Something in between reads each row, turns it into the structure SAP expects, hands it to SAP through an interface, and reads back what SAP said. The eight steps below are that "something in between", whether it is PostNow.ai, a program, a recording or a migration tool.

Four interfaces exist for the hand-over, and the choice shapes steps 4 to 7:

The BAPI vs BDC guide has the detail. For the procedure below, the method decides which tool runs step 5; the other seven steps are the same.

Doing the eight steps in Excel with PostNow.ai

PostNow.ai turns the eight steps below into four things the user does in Excel. The steps are method independent, but the amount of building they need is not: with a program or a recording, steps 3, 4, 6 and 7 are things a developer builds and the business waits for; with PostNow.ai they are clicks in the sheet.

Nothing is installed in SAP, nothing is recorded, and nothing is transported, which is why the same procedure works on ECC and S/4HANA and survives the screen changes that break recordings. The 75 free templates cover the objects most teams upload, with the BAPI structure and field named on every column, so step 2 and step 3 start already done.

The 8 steps to upload Excel data into SAP

PostNow.ai follows these eight steps for every load, and they apply whatever method you use.

1. 1

   Step 1. Define what you are loading, and check you are allowed to

   Name the SAP object and the transaction that creates it, the number of records, the company codes or plants involved, and the date the records must carry. For the worked example: 40 journal entry lines, 10 documents of 4 lines each, one company code, document type SA, posting date the last day of the period.

   Then confirm the SAP user who will run the load has authorisation for that transaction in that organisational unit. An upload never bypasses authorisation; a load run by a user who cannot post to company code 2000 fails on every row for company code 2000, and the failure message is about authorisation, not data. Check it in a five minute manual posting before building anything.

2. 2

   Step 2. Start from a template that names the SAP fields

   PostNow.ai's 75 free templates already name the SAP field in every column. Lay the sheet out with one record per row and the SAP field name, not a description, in the header. BUKRS beats "Company", because the person mapping in step 3 does not have to guess, and because the field name is what every error message will reference.

   For documents with a header and items, journal entries, purchase orders, sales orders, decide the layout now. Two choices work: a flat sheet where the header fields repeat on every item row and a document key column groups them, or a header sheet and an item sheet joined on that key. The flat layout is easier for a user to fill and is what the free templates use.

   A template with SAP field names is the one thing every way to upload Excel to SAP has in common, from an SAP Fiori Excel upload app to a BAPI load.

   For the worked example, the FB50 journal entry template has the header and line columns already named for BAPI_ACC_DOCUMENT_POST: document header (BUKRS, BLDAT, BUDAT, BLART, WAERS), then one line per row with G/L account, debit or credit indicator, amount, cost center and text.

3. 3

   Step 3. Map every column to an SAP field and structure

   PostNow.ai saves the mapping with the template and its Mapping AI proposes the column matches, so this is done once. A mapping is a table with four columns: sheet column, SAP structure, SAP field, rule. The rule is where the work is: a fixed value (document type is always SA), a format conversion (dates as YYYYMMDD), a value translation (the sheet says "Debit", SAP wants S), or a lookup (the cost center by department name).

   Two mappings that catch every first time loader:

   • Leading zeros. SAP stores G/L accounts, cost centers and material numbers with leading zeros to their full length. Excel drops them. 100000 in a cell is 0000100000 in SAP, and a load that sends 100000 either fails or, worse, matches nothing and posts to the wrong account after conversion. Format those columns as text and pad them in the mapping.
   • Amounts and sign. BAPI_ACC_DOCUMENT_POST takes amounts on the CURRENCYAMOUNT table with the sign carrying the debit or credit. A sheet that has a separate debit and credit column needs a rule that produces one signed amount per line.

   If a template from the library is used, this step is a check rather than a build, because the structure and field are in the header already.

4. 4

   Step 4. Validate every row against the live SAP system

   PostNow.ai runs this check for every row against the live system when you click Validate, and points each error at its cell. This is the step that decides whether the load is a load or a clean-up exercise. Validation means checking each row against the target system's configuration and master data before anything posts: the company code exists and is open for the posting period, the G/L account exists in the chart of accounts and is not blocked, the cost center is valid on the posting date and belongs to the company code, the tax code is allowed for the account, the document balances to zero.

   Checking the file's format is not validation. A file can be perfectly formed and 30 percent unpostable because the period was closed yesterday. The check has to be against SAP, and it has to happen before step 5, because a document that fails half way through a batch input session leaves a partial mess that step 7 then has to untangle.

   In the worked example, three of forty lines fail validation: one cost center closed at month end, one G/L account that only accepts postings with a tax code, one amount with a stray text character. All three are fixed in the sheet in two minutes. Found after posting, the same three would have been a reversal, a reposting and an explanation.

   How this step is done depends on the method. With a BAPI, most validation is available through the BAPI's own check mode, BAPI_ACC_DOCUMENT_CHECK for journal entries, or by calling the BAPI with a test run flag where it has one. With a recording, validation is whatever you build. With PostNow.ai, it is the Validate step in the add-in, which runs every row against the live system and writes the SAP message beside each row before the Post button is available.

5. 5

   Step 5. Test in a non-production system first

   PostNow.ai runs the same template against your quality system: connect to it instead of production and run the whole file. Run the same file, with the same mapping, against the quality system. Not a sample of it: the whole file, because the errors that survive validation are the ones that only appear at posting (a substitution rule that changes a field, a user exit that rejects a combination, a number range that has not been maintained in the target client). Compare the result to what you expected: 10 documents, 40 lines, each document balanced, the totals matching the sheet.

   If quality is not a copy of production, and it rarely is, note what differed and re-check those items against production in step 4 before the real run.

6. 6

   Step 6. Post through SAP's own interface

   In PostNow.ai this is the Post button: each validated row posts through the object's BAPI under your own login, and the document number or SAP message is written beside it. Post the validated file to production through the interface chosen in the table above. Three rules hold for every method:

   • One record, one result. Whatever posts must return a result per record: the document number, or the message. A load that returns "session complete" and nothing else has to be reconciled by hand in step 8.
   • Never write to tables. A direct table update skips every check SAP applies, produces documents with no change history, and is the one thing an auditor will not accept. Every method in the table above goes through SAP's business logic; the tool that does not is not a method.
   • Under a named user. The load runs under the identity and authorisation of the person doing it, so that the created-by field on every document is a person and the authorisation check applied is theirs.

   In the worked example the 40 lines become 10 documents, and the sheet gets a column with 10 document numbers, one per line group, written back by the interface.

7. 7

   Step 7. Handle the rows that failed, without double posting

   PostNow.ai writes the SAP result beside every row, so the failed rows are the ones without a document number: filter them, fix them, and post only those. Some rows fail at posting even after validation. The rule for the retry is the one that prevents the most expensive upload mistake there is: only the failed rows are reposted. A file rerun from the top posts every successful document a second time, and 10 duplicate journal entries take longer to reverse than the original load took to run.

   So: filter to the rows with a message, fix the cause (in the sheet if it was data, in SAP if it was configuration), and post that subset. The result column from step 6 is what makes this possible; rows with a document number are done and are excluded from the retry.

8. 8

   Step 8. Reconcile and keep the log

   PostNow.ai leaves the log in the sheet, with the document number or message beside every row. Prove the load: the count of documents created equals the count expected, the totals in SAP (from FBL3N or the GL line items report) equal the totals in the sheet, and the sheet with the SAP document number and message beside every row is saved as the record of the run. That sheet is the audit trail. It answers "was this posted by the upload or by hand", "who ran it", and "which rows failed and what was done about them" a year later.

Which objects can be uploaded from Excel

PostNow.ai uploads any object with a published interface, and has a free template for the common cases below. Anything with a published interface can be uploaded; the common cases, with the template for each:

Customer and vendor master data in S/4HANA are Business Partners, which have their own load page, and mass changes to existing records follow the mass change procedure, which starts from an extract rather than a blank sheet.

Excel uploads in ECC and S/4HANA: what changes

PostNow.ai works on both, because it posts through BAPIs. The eight steps are the same on both, and so are the BAPIs for most objects, which is why a template built on ECC keeps working after a migration. Four differences matter for the mapping in step 3 and the validation in step 4.

Fiori Excel upload apps. S/4HANA ships Fiori apps that accept an Excel template for a single object, the best known being Upload General Journal Entries, which takes a downloaded template, checks it and posts the journal entries. They cover one object each, with SAP's fixed template, and are a good route when the object has one and the template fits. The eight steps above still apply: the template is step 2, the app's check is step 4, and the log is what the app returns.

Batch input. On ECC, the classic SAP batch input Excel route is a recording in SHDB or an LSMW project fed from the sheet. In S/4HANA those recordings break wherever a screen changed, which is why BAPI based loads are the route that carries over.

Fiori changes the screens, not the interfaces. A load that posts through a BAPI does not notice that the transaction now has an app; a recording does. The ECC to S/4HANA migration guide covers the wider changes.

The mistakes that break Excel uploads

PostNow.ai validates every row against the live system before posting, which is where most of these surface. Seven, in the order they usually appear:

1. Leading zeros dropped by Excel on account, cost center and material columns. Format as text before entering data, or pad in the mapping.
2. Dates in the regional format of the PC rather than the format the interface expects. Map to YYYYMMDD explicitly.
3. Descriptions instead of codes. "Raw material" is not a material type; ROH is. Every coded field needs the code, and the mapping holds the translation.
4. Validating the file, not the row against SAP. The period closed, the cost center expired, the account is blocked: none of these are visible in the file.
5. Posting the whole batch on retry. Duplicate documents. Filter to failed rows.
6. Running under a technical user so that every document says it was created by BATCHUSER. Run under the person.
7. No log. A load with no per-row result cannot be reconciled or audited and has to be redone from SAP's side.

Frequently asked questions