Structured drafts
Jianlai tutorial 22: type a CSV source record with commas and escaped quotes
Use one original three-field CSV exercise to record Jianlai character entry, any host rewriting and external field interpretation separately. Retain the raw line and a manual field ledger. This lesson does not treat the keyboard as a CSV importer or confuse separator commas with commas inside fields.
On this page
Use one original three-field CSV exercise to record Jianlai character entry, any host rewriting and external field interpretation separately. Retain the raw line and a manual field ledger. This lesson does not treat the keyboard as a CSV importer or confuse separator commas with commas inside fields.
1. Define the three original field values
On paper, define the first field as an English location phrase containing a comma, the second as a letter combination with one double quote between the letters, and the third as a single character that looks numeric. The raw line below is the sole entry target. It has no header and contains no real customer or account data. Keep field values separate from their file representation: one quote in a value does not mean only one quote is written in the raw record. In this exercise the third field is a textual field value; its appearance as two does not guarantee a numeric type after import. A receiving program may infer a type separately. That is its behavior, not evidence that Jianlai created a spreadsheet.
2. Enter the exact ASCII punctuation
Open a private ordinary text draft, confirm Jianlai is active, and use the installed interface to reach English letters, commas and ASCII double quotes. Enter the opening quote of the first field, its location text and closing quote before the first separator comma. The second field is especially easy to underquote; pause and inspect positions before continuing. The target contains no spaces, final period or surrounding code fences. The article’s fences only identify the sample. Commit composition and inspect the body text. If the host replaces straight quotes with curly quotes, restore the exact target in this copy or use an ordinary editor that preserves it, and record which component performed the substitution.
"north,desk","A""B",2
3. Count the raw characters
Count from left to right. The whole line must contain six ASCII double quotes and three commas. One comma is inside the quoted first field, while the other two separate the fields, so field count is not simply the raw comma count plus one. The first field has one opening and one closing quote. The second has an outer opening and closing quote plus two adjacent quotes representing one quote inside its value: four in total. Draw those roles on paper instead of deleting one of the adjacent straight quotes. Ignore decorative theme borders and inspect committed body characters. Retain the unmodified line as a reference. Seeing what looks like one pair of wide quotation marks is not a substitute for counting the actual characters.
4. Read the three fields back manually
Read the fields back with the ledger below. Keep the internal comma in the first value, decode the adjacent pair of quotes into one quote in the second, and retain the third character unchanged. Ledger labels are explanatory and must not be appended to the raw CSV line. Under the convention described by RFC 4180, fields containing commas or double quotes use outer double quotes, and a quote inside a field is written twice. RFC 4180 is a widely used format description, not a guarantee that every tool called a CSV tool has identical settings. If an external tool defaults to semicolons, record the dialect difference rather than casually rewriting an already verified record to fit the wrong configuration.
field 1: north,desk
field 2: A"B
field 3: 2
5. Make a fault that can be located
Save a separate practice copy and delete only one of the adjacent internal quotes in the second field. Leave the reference line untouched. Explain in the manual ledger that the change breaks the escaped representation of the internal quote, then compare the faulty line with the target character by character to locate the omission. Do not claim every tool reports the same error at the same position. Use the host insertion point or selection to restore the quote precisely and verify that the total returns to six. No upload or repeated production import is needed for this counterexample. If the entire field is selected, cancel the selection and locate the position again so a quoting problem does not become lost content. Retain local before-and-after copies.
6. Optionally check interpretation externally
If you already have a trusted local CSV reader, give it the nonprivate correct sample, specifying comma separation, the quoting convention above and no header. Check for exactly three fields with the listed original values. If it shows two records, first check whether the explanatory ledger or trailing article text was copied too. If it converts the last value to a number, record type inference and display separately; do not describe it as Jianlai changing the source characters. Without a suitable tool, stop at manual reading and mark “external parsing not run.” The paper field expectation does not replace an execution record and does not establish CSV parsing, validation or export as a Jianlai feature.
7. Diagnose extra fields, missing fields and spaces
If the first field splits into two columns, inspect missing outer straight quotes, curly-quote substitutions and whether the reader recognizes quotes under its chosen settings. For a malformed second field, count its adjacent internal quotes. A single-column result can also come from a different delimiter setting; reinstalling the keyboard is not the first diagnostic step. If field values acquire spaces, compare the raw text for habitual spaces inserted after commas. Under this convention spaces are part of a field, so do not assume the receiver ignores them. Fix one discrepancy per round and identify whether the raw line, manual ledger or actual parser report changed. Replacing all punctuation at once can hide the cause rather than diagnose it.
8. Deliver the source, ledger and verification state
Save one clean raw record, a three-field manual ledger and a statement of what was actually checked. Distinguish six double quotes, three raw commas and two separator commas. Close and reopen the host, then copy and compare characters again, recording whether the source survived; do not mark saving as passed if it was not tried. If an external reader was used, list its name, input method and actual output separately from manual expectations. For a longer table later, start with one nonprivate record to reproduce the quoting rule. This lesson completes entry and interpretation of this one record. A correct sample does not establish compatibility with every CSV file, encoding or receiving application.
Related guides and sources
Chinese and ASCII punctuation contracts; Offline address source check.
Checked on 17 November 2026: Apple’s Finland lookup and China lookup both list Jianlai 1.0.5, requiring iOS / iPadOS 16.0 or later. The official store describes themes, backgrounds, fonts and multilingual input. Use controls actually present in the installed version; system editing and host processing are distinguished separately. These are repeatable instructions and paper expectations, not claims of completed device input, host compatibility or format-parser tests.
Method references: RFC 4180 section 2; Apple text editing guide.