Field guide
Combine point files with a verifiable handoff
A useful combined table explains the inputs, the decisions and the resulting records.
Combine records without deciding that two observations are one point
Use this tool to merge survey point files from separate crews, days or deliverable folders into a single reviewable table. It keeps a source identity for each observation, even when two filenames or point numbers are the same. The purpose is a controlled file handoff: align the meaning of columns, account for every record, and make numbering decisions visible before a new file is created.
A merged table is not an adjusted survey network. This page does not average repeated observations, select a more accurate coordinate, register point clouds or join CAD entities. If two records describe the same physical control point, deciding which observation is suitable remains a separate review. Keeping both under different identifiers is often the appropriate way to preserve evidence while that review is unfinished.
Read and confirm each source separately
Choose a UTF-8 CSV or TXT file, inspect its preview, map the columns and confirm the horizontal and elevation units. Read the file and add that interpretation to the queue. Repeat for the next source. Each queued entry is an independent snapshot: selecting another file does not replace the entry already queued. Up to twenty files can be included within the combined size and record limits. XLSX workbooks are also accepted for this input: explicitly choose the worksheet, and replace formulas, dates or error cells with plain values before importing.
Files do not need the same incoming column order. PNEZD places northing before easting; PENZD reverses those two fields. A CASS-style code file can put the description immediately after the identifier. Use explicit mappings rather than assuming that the second column is always a coordinate. To correct a queued interpretation, remove that source and add it again after checking the mapping.
Confirm a common reference before making one table
Matching unit labels do not establish a shared coordinate system. Check the project grid, coordinate origin, height datum and any previous transformations in the source documentation. Enter that common reference in the rule form and confirm it. The tool rejects differing horizontal or elevation unit labels; it does not silently convert one source to match another.
If one crew exported feet and another exported metres, prepare and verify a deliberate conversion first. If the coordinates use different grids, resolve that through the appropriate surveyed transformation workflow. A file merger cannot infer an origin shift, projection or vertical datum from plausible-looking numbers. The reference note is retained in the complete report so the recipient can see the assumption under which the files were assembled.
Keep optional values empty and preserve extra columns
The output begins with Point, Northing, Easting, Elevation and Description. Values are taken from your mappings without rounding coordinates or rewriting codes. An unavailable elevation or description stays empty. The preview counts missing optional values by source and requires acknowledgment before output can be applied. An empty point identifier requires per-source numbering so the final merged table has usable identifiers.
Unmapped columns are not discarded. Each receives a source-qualified heading containing the queue key, original column position and heading where available. This keeps an accuracy field from one file separate from an unrelated field with the same heading in another. Blank cells in those additional columns mean that the record came from a different source, not that a measured attribute was replaced with zero.
Stop on repeated point numbers before anything is overwritten
The initial policy preserves identifiers and blocks application when final IDs conflict. Review the listed source keys and original lines. The report includes every conflict even when the on-screen table shows only a preview. A later file never wins merely because it was imported last, and no record is replaced by a more recently read record.
By default, digit-only IDs are compared numerically, so 001 and 1 conflict. This conservative choice helps when the receiving program treats point numbers as integers. Exact-text comparison is also available for workflows that deliberately distinguish those labels. Letter IDs remain text in either mode. Neither comparison rule changes the original source strings stored in the report.
Keep every record with separate source numbering
Choose the keep-all numbering policy when overlapping number ranges represent separate observations. Each source has its own prefix and starting number in the queue. For example, start the first file at A-001 and the second at B-001 by setting their prefixes to A- and B-. Numbering follows each file's original record order, and the width of a starting value such as 001 is retained.
The complete final set is checked for conflicts, including collisions caused by your chosen prefixes. Two identical prefix/start combinations can still collide and will block application. Renumbering changes only the mapped identifier in the standardized output. It cannot reduce the record count, alter a coordinate or make an uncertain observation valid. Save the old-to-new mapping alongside the new point file.
Exclude only a record you can identify and explain
When a redundant record has been independently confirmed, choose explicit exclusion mode and enter its source key from the preview, such as F2:row-2. This key identifies one parsed record in one queued file. It is not a point-number pattern and does not select every observation bearing the same number. Provide the reason for the exclusion and generate the preview again.
Unknown source keys are rejected rather than ignored. All records not explicitly listed remain included, and any unresolved final-ID conflicts still block application. The complete report retains excluded records and the stated reason. If the conflict remains uncertain, cancel the preview or keep both observations under distinct identifiers instead of guessing which coordinate should survive.
Choose ordering without losing provenance
The default output follows queue order and then original source order. Move a queued file up or down before generating the preview to change that sequence. Alternatively, choose text order by the final identifier. Text sorting is intentionally literal: a name such as 10 can appear before 2. It is not an attempt to infer a natural field-observation sequence.
The mapping records both source start and end lines as well as output line ranges. These differ from record counts when a properly quoted description contains a newline. Two sources can share the same filename without becoming ambiguous because their queue keys are separate. Reordering changes where a record is written, while its source identity remains available for checking.
Reconcile counts with a small example
Suppose Crew A provides one hundred observations and Crew B provides another hundred. With no explicit exclusions, the output must contain two hundred records, regardless of how many numbers collide. The preserve-ID policy may block the output, but the preview still accounts for all two hundred. A successful per-source renumbering also produces two hundred records.
If a reviewer explicitly excludes one confirmed duplicate observation from Crew B, the source summaries show one hundred retained from A and ninety-nine retained plus one excluded from B. The identity input equals retained plus excluded is checked per source. This is a count check, not a statement that the remaining coordinates meet project tolerances. Run the checker and review the point distribution after the merge.
Download a point file and its handoff evidence
After reviewing the full preview, confirm and apply it to enable point CSV and mapping downloads. The point file contains literal source values in a standardized comma-delimited layout with a header. It is intended as data for a surveying workflow. The mapping CSV is a separate review document and protects text that could be interpreted as a spreadsheet formula.
The complete JSON can also preserve an unresolved preview, clearly marked as not applied. It contains original texts, mappings, source summaries, conflicts, excluded records, rule choices and the resulting table. Keep it when handing the merged file to someone else. The next-tool button sends confirmed output and its mapping locally to the checker; the full JSON remains the detailed source trail.
Handle limits and changes predictably
The batch limit is one hundred thousand records and ten MiB of combined input, with at most twenty sources. The serialized output is also limited to ten MiB and 256 columns. Many source-specific extra columns can expand a small input substantially, so an output can reach its limit before the record ceiling. Split the batch or prepare a deliberate reduced schema instead of assuming unlimited capacity.
Long processing runs can be cancelled while retaining the input and queued files. Changing a source, numbering rule or confirmation invalidates the previous preview and disables its point-file downloads. Data is processed in the browser and is not automatically persisted. Before closing the page, download the evidence and output you need; reopening the site does not restore the unsaved queue.