How can I get missing values recorded as NULL when importing from csv - csv

Question

I have multiple, large, csv files, each of which has missing values in many places. When I import the csv file into SQLite, I would like to have the missing values recorded as NULL for the reason that another application expects missing data to be indicated by NULL. My current method does not produce the desired result.
An example CSV file (test.csv) is:
12|gamma|17|delta
67||19|zeta
96|eta||theta
98|iota|29|
The first line is complete; each of the other lines has (or is meant to show!) a single missing item. When I import using:
.headers on
.mode column
.nullvalue NULL
CREATE TABLE t (
id1 INTEGER PRIMARY KEY,
a1 TEXT,
n1 INTEGER,
a2 TEXT
);
.import test.csv t
SELECT
id1, typeof(id1),
a1, typeof(a1),
n1, typeof(n1),
a2, typeof(a2)
FROM t;
the result is
id1 typeof(id1) a1 typeof(a1) n1 typeof(n1) a2 typeof(a2)
---- ----------- ------ ---------- -- ---------- ------ ----------
12 integer gamma text 17 integer delta text
67 integer text 19 integer zeta text
96 integer eta text text theta text
98 integer iota text 29 integer text
so the missing values have become text. I would appreciate some guidance on how to ensure that all missing values become NULL.

Answer 1

sqlite3 imports values as text and there does not seem to be a way to make it treat empty values as nulls.
However, you can update the tables yourself after import, setting empty strings to nulls, like
UPDATE t SET a1=NULL WHERE a1='';
Repeat for each column.
You can also create a trigger for such updates:
CREATE TRIGGER trig_a1 AFTER INSERT ON t WHEN new.a1='' BEGIN
UPDATE t SET a1=NULL WHERE rowid=new.rowid;
END;

Answer 2

For the cases where you cannot update after import because the import will fail when the empty string (text columns) or 0 (integer columns) is inserted instead of NULL, see my answer to this other stackoverflow question