How does browser CSV merging unify mismatched delimiters and encodings?
Dropped .csv, .tsv, and .xlsx files transfer straight into a local Web Worker that inspects every file independently by byte signature, strips leading byte order marks, decodes UTF-8 or Windows-1252 text, and sniffs delimiters before any tables combine. Because processing stays strictly in browser memory, multi-file joins and merges finish with zero server uploads.
Each input table passes through three normalization steps based on the IETF RFC 4180 CSV specification:
- Format and ZIP sniffing: Files carrying a
.xlsxextension or0x50 0x4BZIP header extract their first worksheet directly instead of failing as garbled binary text. - Per-file text decoding: Text tables drop leading BOM bytes and decode strict UTF-8 first, falling back cleanly to legacy Windows-1252 exports when older software drops accented text.
- Delimiter auto-detection: Files ending in
.tsvforce tab separation, while.csvfiles inspect up to10,000characters across10logical rows to select commas, semicolons, tabs, or pipes automatically.
Teams preparing to merge CSV exports alongside CRM directories frequently clean phone or email cards through our VCF to CSV and Excel contact converter first, then combine normalized sheets here with consistent headers.
Stacking tables with union append by header or column position
When combining monthly reports or regional exports that share rows of the same entity type, you choose between aligning fields by column title or stacking raw columns strictly by index. At least two tables are required before merging runs.
Header alignment, blank fallbacks, and _source_file tracking
Choosing Union by matching headers scans headers across every uploaded table in file order, maps empty header cells sequentially to fallback labels (Column 1, Column 2, and so forth), and builds a unified column list. Every row lands under its matching column name, leaving missing columns blank rather than corrupting alignment.
Checking Add _source_file column inserts the original filename into column one so downstream pivot tables keep monthly provenance clear. Per the Unicode UTF-8 BOM standard, stripping input BOM bytes up front prevents invisible 0xEF 0xBB 0xBF markers from creating duplicate first-column header names across files.
Positional stacking and full-row deduplication tradeoffs
Selecting Union by column position aligns cells strictly by column index (0 to maxCols - 1), pulling headers from the first uploaded file and padding shorter files with blank trailing cells. This mode helps when identical vendor schemas ship with slightly translated header strings.
Toggling Remove exact duplicate rows runs full-row signature hashing after stacking completes to deduplicate overlapping entries while reporting the exact count removed in the preview banner. Keep in mind that enabling _source_file tags rows with distinct filenames, preventing identical data across two files from colliding as duplicates.
| Merge Mode | Alignment Rule | Duplicate Column Handling | Missing Field / Key Behavior | Best Use Case |
|---|---|---|---|---|
Union by matching headers |
Header text match | Merges identical column names | Leaves missing headers blank | Monthly exports with shifted columns |
Union by column position |
Column index 0..N |
Keeps first file header | Pads short rows with empty strings | Identical layouts with renamed titles |
Inner join on key column |
Case-insensitive or exact key | Suffixes right duplicates (2) |
Drops unmatched rows from both sides | Finding records present in all files |
Left join on key column |
Sequential left-base match | Suffixes right duplicates (2) |
Retains left rows with empty right cells | Enriching master orders or SKU lists |
Full outer join on key column |
Complete two-sided match | Suffixes right duplicates (2) |
Keeps unmatched rows from either side | Reconciling incomplete customer exports |
When a consolidated file exceeds an upload portal row cap, pass the finished export into a splitter to produce smaller CSV chunks ready for portal row limits without manually slicing ranges in Excel.
Joining CSV and Excel sheets on a matching key column
When you need horizontal enrichment rather than vertical stacking, switch to a key-based join mode and specify a shared identifier such as OrderID or SKU. Loading the built-in sample files (orders-jan.csv and orders-feb.csv) illustrates both overlapping IDs (ORD-1001) and reordered columns right away.
Inner, Left, and Full Outer join behavior across multiple files
Key joins execute sequentially left-to-right across two or more uploaded files, trimming key whitespace and matching case-insensitively by default (or strictly when unchecked). Right-hand rows with empty join keys are skipped so blank cells never cross-join into Cartesian bloat, while valid one-to-many matches expand naturally.
To avoid erasing data during multi-table joins, colliding column names from joined files automatically gain numbered suffixes such as (2) or (3). Native .xlsx sheets parsed under the ECMA-376 OpenXML workbook specification read their first tab directly alongside .csv or .tsv lookups.
Validating joined spreadsheets and auditing row edits
After executing a Left or Full Outer join, preview up to 100 combined rows directly in the grid alongside exact dimension totals before downloading. When validating updated pricing or order exports before pushing changes live, it helps to compare before and after spreadsheets for cell edits side by side.
Exporting merged CSV, TSV, and XLSX with formula neutralization
Once your preview looks accurate, choose Export merged CSV, TSV, or XLSX to download merged-tables.csv, merged-tables.tsv, or merged-tables.xlsx covering every row beyond the 100-row browser preview cap. Embedded OpenXML parsing protects browser stability with hard safety caps of 180 MB total unpacked workbook XML, 100 MB per compressed ZIP entry, and 250,000 rows per worksheet.
- Merged Excel (.xlsx): Packs output into a single
Mergedworksheet usinginlineStrcells, bold Calibri 11 pt row-one headers, and auto-fitted column widths clamped between8and60characters. - CSV and TSV safety: Prepends a UTF-8 BOM (
0xEF 0xBB 0xBF) for reliable Excel opening and applies MITRE CWE-1236 formula injection guidance by default whenever Escape formula prefixes stays checked.
Formula neutralization prefixes leading =, +, @, |, and non-numeric - strings with a single apostrophe so untrusted cells from merged web forms cannot execute formulas. If you frequently need formatted Excel workbooks ready for sharing after merging plain delimited tables, keep that companion formatter bookmarked alongside this tool.