{ "slug": "csv-null-standardizer", "name": "gizza-ai/csv-null-standardizer", "version": "0.1.0", "title": "CSV Null Standardizer — One Missing-Value Token — gizza.ai", "description": "Paste a CSV or TSV and rewrite NA, N/A, #N/A, NULL, NaN, None, dashes and blank cells into one missing-value token. Runs locally in your browser.", "tags": [ "csv null standardizer", "missing values", "na values", "null representation", "normalize nulls", "csv cleaning", "na_rep", "data quality" ], "category": "data", "urls": { "page": "https://gizza.ai/tools/csv-null-standardizer/", "markdown": "https://gizza.ai/tools/csv-null-standardizer/index.md", "descriptor": "https://gizza.ai/tools/csv-null-standardizer/tool.json", "deep_link_example": "https://gizza.ai/tools/csv-null-standardizer/?input=id%2Cscore%2Cnotes%0A1%2C42%2Cok%0A2%2CNA%2C%0A3%2Cnull%2C-%0A4%2Cn%2Fa%2CNaN&delimiter=comma&na_tokens=NA%2CN%2FA%2CN.A.%2C%23N%2FA%2C%23N%2FA%20N%2FA%2C%23NA%2CNULL%2CNIL%2CNaN%2CNone%2C%3CNA%3E%2C-%2C--%2C%3F&replace_with=NULL&blank_is_missing=true&case_sensitive=true&trim=true&header=true&columns=score%2Cnotes"e_style=minimal" }, "cli": "gizza tool csv-null-standardizer \"id,score,notes\n1,42,ok\n2,NA,\n3,null,-\n4,n/a,NaN\"", "tool": { "description": "Standardize missing-value tokens in a CSV/delimited table. Real-world exports mix NA, N/A, #N/A, NULL, NaN, None, , a bare dash and empty cells in the same file; this rewrites every one of them into a single representation you choose, so a loader or dataframe needs one rule instead of a dozen. Non-missing cells are copied verbatim (no trimming, no re-typing), the header row is never rewritten, and the field separator round-trips unchanged. delimiter accepts 'auto' (sniffed from the first line), a single char, or 'comma'/'tab'/'semicolon'/'pipe'. na_tokens is the comma-separated missing vocabulary (default 'NA,N/A,N.A.,#N/A,#N/A N/A,#NA,NULL,NIL,NaN,None,,-,--,?'; add sentinels like '-999'; empty = blanks only). replace_with is what every missing cell becomes (default empty = a blank cell; try 'NULL', 'NaN', or '\\N' for Postgres COPY). blank_is_missing (default on) includes empty/whitespace-only cells; case_sensitive (default off) requires exact case; trim (default on) ignores padding while matching; header (default on) protects row 1; columns limits the rewrite to named or 1-based columns; quote_style is 'minimal', 'always', or 'never'. This normalizes the TOKEN — it never invents a value, so use an imputer if you need means/medians filled in. Runs entirely in the sandbox; nothing is uploaded.", "parameters": { "additionalProperties": false, "properties": { "blank_is_missing": { "default": true, "description": "When true (default), a cell that is empty or whitespace-only also counts as missing and is rewritten. Turn it off to convert only the listed tokens and leave already-blank cells alone.", "type": "boolean" }, "case_sensitive": { "default": false, "description": "When false (default), tokens match regardless of case, so one 'NULL' entry also catches 'null' and 'Null'. Turn it on to require an exact-case match.", "type": "boolean" }, "columns": { "default": "", "description": "Restrict the rewrite to specific columns: a comma-separated list of column names (needs header = true) or 1-based positions, e.g. 'score,notes' or '2,4'. Default (empty) standardizes every column.", "type": "string" }, "delimiter": { "default": ",", "description": "Field separator: 'auto' to sniff it from the first line, a single character, or a name ('comma' (default), 'tab', 'semicolon', 'pipe'). The output uses the same separator as the input.", "type": "string" }, "header": { "default": true, "description": "When true (default), the first row is a header: it is copied through untouched, so a column literally named 'NA' survives, and it supplies the names used by the columns option.", "type": "boolean" }, "input": { "description": "The CSV/delimited table to standardize, as text. Quoted fields with embedded separators or newlines (RFC 4180) are preserved. Rows may be ragged — a short row stays short. Max 5,000,000 bytes.", "type": "string" }, "na_tokens": { "default": "NA,N/A,N.A.,#N/A,#N/A N/A,#NA,NULL,NIL,NaN,None,,-,--,?", "description": "Comma-separated tokens that count as missing, e.g. 'NA,N/A,NULL,NaN,-'. Defaults to the built-in vocabulary 'NA,N/A,N.A.,#N/A,#N/A N/A,#NA,NULL,NIL,NaN,None,,-,--,?'. Edit the list to add your own sentinels (e.g. '-999'). Set it empty to standardize ONLY blank cells.", "type": "string" }, "quote_style": { "default": "minimal", "description": "Output quoting: 'minimal' (default) quotes only fields that need it; 'always' quotes every field, the replacement token included, for readers that expect uniformly quoted CSV; 'never' emits bare fields, which is compact but can produce ambiguous CSV when values contain the separator. Keep 'minimal' when replace_with is a loader sentinel such as '\\N', because a quoted sentinel is read as a literal string.", "enum": [ "minimal", "always", "never" ], "type": "string" }, "replace_with": { "default": "", "description": "The single representation written for every missing cell, e.g. 'NULL', 'NA', 'NaN', or '\\N' for a Postgres COPY load. Default is an empty string, which leaves the cell blank.", "type": "string" }, "trim": { "default": true, "description": "When true (default), whitespace around a cell is ignored while matching, so ' NA ' still counts as missing. Cells that are NOT missing are always copied verbatim, padding included — this option never rewrites real values.", "type": "boolean" } }, "required": [ "input" ], "type": "object" } } }