Difference between revisions of "MakeCSV"
|Line 21:||Line 21:|
=== How values are quoted ===
=== How values are quoted ===
, can the . <code></code> <code></code> <code></code> <code>=</code>=
required the text contains a separator character ,
a quote appears text, it example text <code>""" "</code>.
=== dateTemplate ===
=== dateTemplate ===
Revision as of 17:23, 4 September 2019
new to Analytica 5.0
- 1 MakeCSV(a, columnIndex, rowIndex, columnHeaders, rowHeaders, separator, quoteFlags, dateTemplate, decimal, quote)
- 2 Examples
- 3 See Also
MakeCSV(a, columnIndex, rowIndex, columnHeaders, rowHeaders, separator, quoteFlags, dateTemplate, decimal, quote)
Makes CSV-formatted text from a two-dimensional array «a» which is indexed by «columnIndex» and «rowIndex». CSV stands for "Comma separated value", but may use a separator other than a comma in practice. Each row is separated in the result by a new-line (CR), and within a row each item is separated by a comma by default, or by «separator».
Null values are blank, with no space between the leading and separators.
Row and Column Headers
A first line can be optionally included containing the column names. If you want a column header line, specify the names of the columns to «columnHeaders». The value you pass must be indexed by «columnIndex». When your column index contains the labels you want to use, simple pass that same index to «columnHeaders». If you omit «columnHeaders», the first line contains the first row of data.
You can optionally prepend a first column containing row headers. To do so, specify the row header labels or values in the «rowHeaders» parameter. The value you pass must be indexed by «rowIndex». When your row index already contains the row headers, pass the same row index for «rowHeaders».
Changing the field separator
To use a separator other than comma, specify it in the «separator» parameter. «Separator» can contain multiple characters.
A commonly used separator is the tab-character, which you can specify as
Chr(9), or you can type a tab character in a definition with the three keystrokes quote-tab-quote.
In some countries and regions that use comma as the decimal point for numbers, Excel does not recognize the comma as a separator when you try to read in a CSV file. You may need to specify a semi-colon separator in these cases, or the Excel user can read in the CSV file and then convert it from text to columns in a separate step after he loads it into Excel.
How values are quoted
By default, it puts quotes around text and no quotes around numbers and dates. For a text value that can parse as a number or date, it puts an "=" sign before the text. For example,
="1/2/3". This follows a quirk of Microsoft Excel, which, without a "=", converts such text strings into a number or date.
You can override these defaults with the «quoteFlags» parameter. It is a bit field so you can add any of these flags together to combine their effects:
1: No quotes around text (except when required because the text contains a field separator character (default ',') or newsline character)
2: No "=" before a text that parses as number
4: No "=" before a text that parses as date
8: Put quotes around numbers
16: Put quotes around dates
If a quote appears inside quoted text, it doubles the quote. For example text
"Say "Hello" "becomes
"Say ""Hello"" ". Microsoft Excel and Analytica treat the double quote as a single quote inside the text string when parsing it.
The «dateTemplate» controls how date-time numbers are written. To control the format of date-time numbers, you can specify a template using the same format used for a custom date format in the Number Format Dialog. For example, a «dateTemplate» of "dd/MM/yyyy" would write
The default date template is "SHORT¿ TIME?", which is the regional short date format when there is no time component (e.g.,
5/13/2016 in the US,
13/05/2016 in Argentina) or the regional short date format followed by the long time format when there is a time component (e.g,
5/13/2016 03:38:01 p.m.). Regional formats are set in the Region and Language dialog in the Windows Control Panel.
When there is both a date and a time component, experimentation has shown that Excel reads in the date part, but drops the time part. This seems to be an inherent limitation of Excel -- we have not found any date-time format or combination of quoting that works around this limitation.
Changing the decimal point character
The «decimal» parameter specifies the character used to separate the whole number from the fractional part of a number. When not specified, this defaults to the standard convention in the United States of using a dot, e.g.,
3.14. To use a comma, use
MakeCSV(..., separator:';', decimal:','), which writes
3,14. MakeCSV does not write thousand separators within numbers. When you change the «decimal» to comma, it is recommended that you use a «separator» other than comma, such as a semi-colon.
By default, the double quote character (
") is used. You can specify a different character to use in the «quote» parameter. For example, to use single quotes, use
The value here is typed as four single quotes, which is one way to type a text literal in Analytica consisting of a single character, which happens to be the single quote character. (Another way to type it is doubleQuote - singleQuote - doubleQuote).
Create a text value in standard CSV form from an array
Salary indexed by
MakeCSV(Salary, Time, Person)
This puts each person on a separate line, and the columns across correspond to Time. There are so row or column headers, so the first line of data is the first person. Here is a sampling (with 5 Time points and 3 Persons):
45600, 46500, 48800, 51000, 52000 ,,71000, 73000, 77500 120000, 120000, 125000, 125000, 125000
To write the same data to a csv file, use
WriteTextFile("salary data.csv", MakeCSV(Salary, Time, Person))
To include the person as the first field on each line, and the Time for each column on the first line, use
MakeCSV(Salary, Time, Person, columnHeaders: Time, rowHeaders: Person) →
"Person", 2011, 2012, 2013, 2014, 2015 "00001", 45600, 46500, 48800, 51000, 52000 "00002",,, 71000, 73000, 77500 "00003", 120000, 120000, 125000, 125000, 125000
In the CSV here, the
Person index contains textual ID numbers with leading zeros.
LastName are indexed by
Person, full names can be used for the row header by using
MakeCSV(Salary, Time, Person, columnHeaders: Time, rowHeaders: FirstName & ' ' & LastName) →
"Person", 2011, 2012, 2013, 2014, 2015 "Alice Zen", 45600, 46500, 48800, 51000, 52000 "Bob Yarbough",,, 71000, 73000, 77500 "Carol Xian", 120000, 120000, 125000, 125000, 125000