Detect blank or placeholder values
Exacttext_nulls_classifyDeterministic and authoritative — the canonical answer for this input.
Classify whether a value represents 'missing' in messy real-world data. Detects: empty strings, common null markers (NA, N/A, UNK, NONE, NULL, NIL, TBD, NOT APPLICABLE, -, --, .), and Excel error strings (#VALUE!, #REF!, #N/A, #NAME?, #DIV/0!, #NULL!, #NUM!). Accepts any JSON type (strings, numbers, None). Returns `{is_null, reason, matched_token}`. Use this as a pre-clean step before calling other Snipget tools on user-provided data. Processing multiple inputs? Use `text_nulls_classify_batch` to do up to 1000 in one call.
Parameters
valuestringrequiredValue to classify. Accepts any JSON type.
Call it
curl -X POST https://api.snipget.ai/v1/text/nulls/classify \
-H "Authorization: Bearer sk_live_..." \
-H "Content-Type: application/json" \
-d '{"value":"N/A"}'The Batch tab sends up to 1000 inputs in one call via /v1/text/nulls/classify/batch.
Example response
{
"is_null": true,
"reason": "null_marker",
"matched_token": "N/A"
}FAQ
Why not just check for an empty cell myself?
Real-world data hides 'missing' in many disguises: NA, N/A, UNK, NONE, NULL, TBD, a lone dash, or Excel errors like #REF! and #N/A. This recognizes all of those in one consistent check, so you do not have to remember every variant.
What exactly counts as missing?
Empty or whitespace-only strings, common null markers (NA, N/A, UNK, UNKNOWN, NONE, NULL, NIL, TBD, NOT APPLICABLE, -, --, .), and Excel error strings (#VALUE!, #REF!, #N/A, #NAME?, #DIV/0!, #NULL!, #NUM!). The check is case-insensitive.
Does it treat 0 or False as missing?
No. The number zero and the value false are real, legitimate values, so they come back as not null. Only true placeholders are flagged.
What does the result tell me?
Three fields: is_null (true or false), reason (empty, null_marker, excel_error, or non_null), and matched_token (which marker it matched, or null). Use it as a pre-clean step before running other tools.
Is the answer stable?
Yes. It is deterministic and authoritative: the same value always classifies the same way.